commit 5c5a1eb5c0a2d3a503e8525549a3ef6d2d3dc931 Author: ModelHub XC Date: Wed Sep 16 02:10:12 2026 +0800 初始化项目,由ModelHub XC社区提供模型 Model: laion/exp_tas_linear_history_off_traces Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..197e23f --- /dev/null +++ b/.gitattributes @@ -0,0 +1,56 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text + + +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text + +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.gguf* filter=lfs diff=lfs merge=lfs -text +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text + +training_args.bin filter=lfs diff=lfs merge=lfs -text +model-00001-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00004-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +merges.txt filter=lfs diff=lfs merge=lfs -text +vocab.json filter=lfs diff=lfs merge=lfs -text +model-00003-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00002-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 0000000..8c92bb3 --- /dev/null +++ b/README.md @@ -0,0 +1,61 @@ +--- +library_name: transformers +license: apache-2.0 +base_model: Qwen/Qwen3-8B +tags: +- llama-factory +- full +- generated_from_trainer +model-index: +- name: exp_tas_linear_history_off_traces + results: [] +--- + + + +# exp_tas_linear_history_off_traces + +This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the DCAgent/exp_tas_linear_history_off_traces dataset. + +## Model description + +More information needed + +## Intended uses & limitations + +More information needed + +## Training and evaluation data + +More information needed + +## Training procedure + +### Training hyperparameters + +The following hyperparameters were used during training: +- learning_rate: 4e-05 +- train_batch_size: 1 +- eval_batch_size: 8 +- seed: 42 +- distributed_type: multi-GPU +- num_devices: 8 +- gradient_accumulation_steps: 2 +- total_train_batch_size: 16 +- total_eval_batch_size: 64 +- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.98) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments +- lr_scheduler_type: cosine +- lr_scheduler_warmup_ratio: 0.1 +- num_epochs: 7.0 + +### Training results + + + +### Framework versions + +- Transformers 4.57.3 +- Pytorch 2.9.0+cu128 +- Datasets 4.4.1 +- Tokenizers 0.22.1 diff --git a/added_tokens.json b/added_tokens.json new file mode 100644 index 0000000..b54f913 --- /dev/null +++ b/added_tokens.json @@ -0,0 +1,28 @@ +{ + "": 151668, + "": 151658, + "": 151666, + "": 151667, + "": 151657, + "": 151665, + "<|box_end|>": 151649, + "<|box_start|>": 151648, + "<|endoftext|>": 151643, + "<|file_sep|>": 151664, + "<|fim_middle|>": 151660, + "<|fim_pad|>": 151662, + "<|fim_prefix|>": 151659, + "<|fim_suffix|>": 151661, + "<|im_end|>": 151645, + "<|im_start|>": 151644, + "<|image_pad|>": 151655, + "<|object_ref_end|>": 151647, + "<|object_ref_start|>": 151646, + "<|quad_end|>": 151651, + "<|quad_start|>": 151650, + "<|repo_name|>": 151663, + "<|video_pad|>": 151656, + "<|vision_end|>": 151653, + "<|vision_pad|>": 151654, + "<|vision_start|>": 151652 +} diff --git a/all_results.json b/all_results.json new file mode 100644 index 0000000..b587b25 --- /dev/null +++ b/all_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 3.6947044912542757, + "achieved_tflops_per_gpu_theoretical": 392.9969292874875, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.312700480222702, + "mfu_percent": 0.3735798272249015, + "mfu_percent_theoretical": 39.736797703487106, + "total_flos": 9.579822550267986e+17, + "train_loss": 0.48628626106136136, + "train_runtime": 32410.6521, + "train_samples_per_second": 1.851, + "train_steps_per_second": 0.116, + "valid_targets_mean": 4984.9, + "valid_targets_min": 1323 +} \ No newline at end of file diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000..01be9b3 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,89 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0].role == 'system' %} + {{- messages[0].content + '\n\n' }} + {%- endif %} + {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0].role == 'system' %} + {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('') and message.content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} +{%- endfor %} +{%- for message in messages %} + {%- if message.content is string %} + {%- set content = message.content %} + {%- else %} + {%- set content = '' %} + {%- endif %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- if loop.index0 > ns.last_query_index %} + {%- if loop.last or (not loop.last and reasoning_content) %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls %} + {%- for tool_call in message.tool_calls %} + {%- if (loop.first and content) or (not loop.first) %} + {{- '\n' }} + {%- endif %} + {%- if tool_call.function %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {%- if tool_call.arguments is string %} + {{- tool_call.arguments }} + {%- else %} + {{- tool_call.arguments | tojson }} + {%- endif %} + {{- '}\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/config.json b/config.json new file mode 100644 index 0000000..f8a98d4 --- /dev/null +++ b/config.json @@ -0,0 +1,68 @@ +{ + "architectures": [ + "Qwen3ForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "dtype": "bfloat16", + "eos_token_id": 151645, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 40960, + "max_window_layers": 36, + "model_type": "qwen3", + "num_attention_heads": 32, + "num_hidden_layers": 36, + "num_key_value_heads": 8, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_scaling": null, + "rope_theta": 1000000, + "sliding_window": null, + "tie_word_embeddings": false, + "transformers_version": "4.57.3", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/configuration.json b/configuration.json new file mode 100644 index 0000000..159097f --- /dev/null +++ b/configuration.json @@ -0,0 +1 @@ +{"framework": "pytorch", "task": "others", "allow_remote": true} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..9047f32 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,12 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "temperature": 0.6, + "top_k": 20, + "top_p": 0.95, + "transformers_version": "4.57.3" +} diff --git a/merges.txt b/merges.txt new file mode 100644 index 0000000..80c1a19 --- /dev/null +++ b/merges.txt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5 +size 1671853 diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors new file mode 100644 index 0000000..0fedbd9 --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:488ef03013b76ea1d2a97a7c749eaab16bb6c257d180bdbdc7c0ec802a09b53e +size 4902257696 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..155ce84 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4e649b8ae3b1e45cd38606401d8b9d283a61d1be2c9bf8261378f9d02d6e33ca +size 4915960368 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..416941e --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:127fd49933e073f2d394aff34dd97f9f0faadbe47bf424117062e60b868da3dc +size 4983068496 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..9f524d5 --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38d7de20968f00844afeace2e795847baaa7d01951123137e5b232a9e7c99e4e +size 1580230264 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000..ba886c0 --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,407 @@ +{ + "metadata": { + "total_parameters": 308224, + "total_size": 16381470720 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.gate_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/run_summary.json b/run_summary.json new file mode 100644 index 0000000..8a7ab60 --- /dev/null +++ b/run_summary.json @@ -0,0 +1,12 @@ +{ + "agent_name": null, + "training_start": null, + "training_end": null, + "created_by": "DCAgent", + "base_model_name": "Qwen/Qwen3-8B", + "dataset_name": "DCAgent/exp_tas_linear_history_off_traces", + "training_type": "SFT", + "training_parameters": "https://huggingface.co/laion/exp_tas_linear_history_off_traces/blob/main/config.json", + "wandb_link": "https://wandb.ai/DCAgent/OpenThoughts-Agent/runs/exp_tas_linear_history_off_traces_Qwen3-8B", + "traces_location_s3": null +} \ No newline at end of file diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..ac23c0a --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,31 @@ +{ + "additional_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "eos_token": { + "content": "<|im_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..cd71f61 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4 +size 11422654 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..e9dc937 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,240 @@ +{ + "add_bos_token": false, + "add_prefix_space": false, + "added_tokens_decoder": { + "151643": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151644": { + "content": "<|im_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151645": { + "content": "<|im_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151646": { + "content": "<|object_ref_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151647": { + "content": "<|object_ref_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151648": { + "content": "<|box_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151649": { + "content": "<|box_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151650": { + "content": "<|quad_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151651": { + "content": "<|quad_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151652": { + "content": "<|vision_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151653": { + "content": "<|vision_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151654": { + "content": "<|vision_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151655": { + "content": "<|image_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151656": { + "content": "<|video_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151657": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151658": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151659": { + "content": "<|fim_prefix|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151660": { + "content": "<|fim_middle|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151661": { + "content": "<|fim_suffix|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151662": { + "content": "<|fim_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151663": { + "content": "<|repo_name|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151664": { + "content": "<|file_sep|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151665": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151666": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151667": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151668": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": {}, + "model_max_length": 32768, + "pad_token": "<|endoftext|>", + "padding_side": "right", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/train_results.json b/train_results.json new file mode 100644 index 0000000..b587b25 --- /dev/null +++ b/train_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 3.6947044912542757, + "achieved_tflops_per_gpu_theoretical": 392.9969292874875, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.312700480222702, + "mfu_percent": 0.3735798272249015, + "mfu_percent_theoretical": 39.736797703487106, + "total_flos": 9.579822550267986e+17, + "train_loss": 0.48628626106136136, + "train_runtime": 32410.6521, + "train_samples_per_second": 1.851, + "train_steps_per_second": 0.116, + "valid_targets_mean": 4984.9, + "valid_targets_min": 1323 +} \ No newline at end of file diff --git a/trainer_log.jsonl b/trainer_log.jsonl new file mode 100644 index 0000000..42122e3 --- /dev/null +++ b/trainer_log.jsonl @@ -0,0 +1,751 @@ +{"current_steps": 5, "total_steps": 3752, "loss": 0.9659, "lr": 4.2553191489361704e-07, "epoch": 0.009328358208955223, "percentage": 0.13, "elapsed_time": "0:00:56", "remaining_time": "11:42:05"} +{"current_steps": 10, "total_steps": 3752, "loss": 0.9484, "lr": 9.574468085106384e-07, "epoch": 0.018656716417910446, "percentage": 0.27, "elapsed_time": "0:01:50", "remaining_time": "11:30:20"} +{"current_steps": 15, "total_steps": 3752, "loss": 0.9448, "lr": 1.4893617021276596e-06, "epoch": 0.027985074626865673, "percentage": 0.4, "elapsed_time": "0:02:41", "remaining_time": "11:08:59"} +{"current_steps": 20, "total_steps": 3752, "loss": 0.9586, "lr": 2.021276595744681e-06, "epoch": 0.03731343283582089, "percentage": 0.53, "elapsed_time": "0:03:14", "remaining_time": "10:03:31"} +{"current_steps": 25, "total_steps": 3752, "loss": 0.9296, "lr": 2.553191489361702e-06, "epoch": 0.04664179104477612, "percentage": 0.67, "elapsed_time": "0:04:05", "remaining_time": "10:10:10"} +{"current_steps": 30, "total_steps": 3752, "loss": 0.8969, "lr": 3.0851063829787237e-06, "epoch": 0.055970149253731345, "percentage": 0.8, "elapsed_time": "0:04:52", "remaining_time": "10:04:27"} +{"current_steps": 35, "total_steps": 3752, "loss": 0.8351, "lr": 3.6170212765957453e-06, "epoch": 0.06529850746268656, "percentage": 0.93, "elapsed_time": "0:05:35", "remaining_time": "9:54:26"} +{"current_steps": 40, "total_steps": 3752, "loss": 0.8389, "lr": 4.148936170212766e-06, "epoch": 0.07462686567164178, "percentage": 1.07, "elapsed_time": "0:06:08", "remaining_time": "9:29:25"} +{"current_steps": 45, "total_steps": 3752, "loss": 0.8262, "lr": 4.680851063829788e-06, "epoch": 0.08395522388059702, "percentage": 1.2, "elapsed_time": "0:06:39", "remaining_time": "9:08:22"} +{"current_steps": 50, "total_steps": 3752, "loss": 0.7966, "lr": 5.212765957446809e-06, "epoch": 0.09328358208955224, "percentage": 1.33, "elapsed_time": "0:07:29", "remaining_time": "9:14:08"} +{"current_steps": 55, "total_steps": 3752, "loss": 0.8038, "lr": 5.744680851063831e-06, "epoch": 0.10261194029850747, "percentage": 1.47, "elapsed_time": "0:08:08", "remaining_time": "9:07:32"} +{"current_steps": 60, "total_steps": 3752, "loss": 0.7915, "lr": 6.276595744680851e-06, "epoch": 0.11194029850746269, "percentage": 1.6, "elapsed_time": "0:08:49", "remaining_time": "9:03:27"} +{"current_steps": 65, "total_steps": 3752, "loss": 0.7856, "lr": 6.808510638297873e-06, "epoch": 0.12126865671641791, "percentage": 1.73, "elapsed_time": "0:09:34", "remaining_time": "9:02:48"} +{"current_steps": 70, "total_steps": 3752, "loss": 0.7586, "lr": 7.340425531914894e-06, "epoch": 0.13059701492537312, "percentage": 1.87, "elapsed_time": "0:10:25", "remaining_time": "9:08:03"} +{"current_steps": 75, "total_steps": 3752, "loss": 0.7278, "lr": 7.872340425531916e-06, "epoch": 0.13992537313432835, "percentage": 2.0, "elapsed_time": "0:11:06", "remaining_time": "9:04:17"} +{"current_steps": 80, "total_steps": 3752, "loss": 0.6994, "lr": 8.404255319148937e-06, "epoch": 0.14925373134328357, "percentage": 2.13, "elapsed_time": "0:11:56", "remaining_time": "9:07:50"} +{"current_steps": 85, "total_steps": 3752, "loss": 0.7249, "lr": 8.936170212765958e-06, "epoch": 0.15858208955223882, "percentage": 2.27, "elapsed_time": "0:12:44", "remaining_time": "9:09:46"} +{"current_steps": 90, "total_steps": 3752, "loss": 0.6834, "lr": 9.46808510638298e-06, "epoch": 0.16791044776119404, "percentage": 2.4, "elapsed_time": "0:13:37", "remaining_time": "9:14:14"} +{"current_steps": 95, "total_steps": 3752, "loss": 0.7944, "lr": 1e-05, "epoch": 0.17723880597014927, "percentage": 2.53, "elapsed_time": "0:14:21", "remaining_time": "9:13:00"} +{"current_steps": 100, "total_steps": 3752, "loss": 0.7016, "lr": 1.0531914893617022e-05, "epoch": 0.1865671641791045, "percentage": 2.67, "elapsed_time": "0:15:04", "remaining_time": "9:10:40"} +{"current_steps": 105, "total_steps": 3752, "loss": 0.7004, "lr": 1.1063829787234044e-05, "epoch": 0.1958955223880597, "percentage": 2.8, "elapsed_time": "0:15:43", "remaining_time": "9:06:25"} +{"current_steps": 110, "total_steps": 3752, "loss": 0.6832, "lr": 1.1595744680851065e-05, "epoch": 0.20522388059701493, "percentage": 2.93, "elapsed_time": "0:16:27", "remaining_time": "9:05:11"} +{"current_steps": 115, "total_steps": 3752, "loss": 0.6627, "lr": 1.2127659574468087e-05, "epoch": 0.21455223880597016, "percentage": 3.07, "elapsed_time": "0:17:11", "remaining_time": "9:03:53"} +{"current_steps": 120, "total_steps": 3752, "loss": 0.706, "lr": 1.2659574468085108e-05, "epoch": 0.22388059701492538, "percentage": 3.2, "elapsed_time": "0:17:50", "remaining_time": "8:59:57"} +{"current_steps": 125, "total_steps": 3752, "loss": 0.6307, "lr": 1.3191489361702127e-05, "epoch": 0.2332089552238806, "percentage": 3.33, "elapsed_time": "0:18:30", "remaining_time": "8:57:05"} +{"current_steps": 130, "total_steps": 3752, "loss": 0.6731, "lr": 1.372340425531915e-05, "epoch": 0.24253731343283583, "percentage": 3.46, "elapsed_time": "0:19:12", "remaining_time": "8:55:01"} +{"current_steps": 135, "total_steps": 3752, "loss": 0.6576, "lr": 1.425531914893617e-05, "epoch": 0.251865671641791, "percentage": 3.6, "elapsed_time": "0:20:02", "remaining_time": "8:56:59"} +{"current_steps": 140, "total_steps": 3752, "loss": 0.6502, "lr": 1.4787234042553193e-05, "epoch": 0.26119402985074625, "percentage": 3.73, "elapsed_time": "0:20:44", "remaining_time": "8:55:07"} +{"current_steps": 145, "total_steps": 3752, "loss": 0.693, "lr": 1.5319148936170214e-05, "epoch": 0.27052238805970147, "percentage": 3.86, "elapsed_time": "0:21:26", "remaining_time": "8:53:29"} +{"current_steps": 150, "total_steps": 3752, "loss": 0.6717, "lr": 1.5851063829787235e-05, "epoch": 0.2798507462686567, "percentage": 4.0, "elapsed_time": "0:22:16", "remaining_time": "8:55:02"} +{"current_steps": 155, "total_steps": 3752, "loss": 0.6199, "lr": 1.6382978723404255e-05, "epoch": 0.2891791044776119, "percentage": 4.13, "elapsed_time": "0:23:15", "remaining_time": "8:59:40"} +{"current_steps": 160, "total_steps": 3752, "loss": 0.6311, "lr": 1.6914893617021276e-05, "epoch": 0.29850746268656714, "percentage": 4.26, "elapsed_time": "0:24:01", "remaining_time": "8:59:23"} +{"current_steps": 165, "total_steps": 3752, "loss": 0.6028, "lr": 1.74468085106383e-05, "epoch": 0.30783582089552236, "percentage": 4.4, "elapsed_time": "0:24:33", "remaining_time": "8:53:51"} +{"current_steps": 170, "total_steps": 3752, "loss": 0.626, "lr": 1.797872340425532e-05, "epoch": 0.31716417910447764, "percentage": 4.53, "elapsed_time": "0:25:17", "remaining_time": "8:52:44"} +{"current_steps": 175, "total_steps": 3752, "loss": 0.663, "lr": 1.8510638297872342e-05, "epoch": 0.32649253731343286, "percentage": 4.66, "elapsed_time": "0:25:49", "remaining_time": "8:47:51"} +{"current_steps": 180, "total_steps": 3752, "loss": 0.5866, "lr": 1.9042553191489363e-05, "epoch": 0.3358208955223881, "percentage": 4.8, "elapsed_time": "0:26:30", "remaining_time": "8:45:57"} +{"current_steps": 185, "total_steps": 3752, "loss": 0.6381, "lr": 1.9574468085106384e-05, "epoch": 0.3451492537313433, "percentage": 4.93, "elapsed_time": "0:27:17", "remaining_time": "8:46:06"} +{"current_steps": 190, "total_steps": 3752, "loss": 0.6347, "lr": 2.0106382978723404e-05, "epoch": 0.35447761194029853, "percentage": 5.06, "elapsed_time": "0:27:59", "remaining_time": "8:44:46"} +{"current_steps": 195, "total_steps": 3752, "loss": 0.6658, "lr": 2.063829787234043e-05, "epoch": 0.36380597014925375, "percentage": 5.2, "elapsed_time": "0:28:38", "remaining_time": "8:42:27"} +{"current_steps": 200, "total_steps": 3752, "loss": 0.6138, "lr": 2.117021276595745e-05, "epoch": 0.373134328358209, "percentage": 5.33, "elapsed_time": "0:29:23", "remaining_time": "8:42:08"} +{"current_steps": 205, "total_steps": 3752, "loss": 0.5572, "lr": 2.1702127659574467e-05, "epoch": 0.3824626865671642, "percentage": 5.46, "elapsed_time": "0:30:00", "remaining_time": "8:39:08"} +{"current_steps": 210, "total_steps": 3752, "loss": 0.6368, "lr": 2.223404255319149e-05, "epoch": 0.3917910447761194, "percentage": 5.6, "elapsed_time": "0:30:40", "remaining_time": "8:37:15"} +{"current_steps": 215, "total_steps": 3752, "loss": 0.6236, "lr": 2.2765957446808512e-05, "epoch": 0.40111940298507465, "percentage": 5.73, "elapsed_time": "0:31:28", "remaining_time": "8:37:54"} +{"current_steps": 220, "total_steps": 3752, "loss": 0.5931, "lr": 2.3297872340425536e-05, "epoch": 0.41044776119402987, "percentage": 5.86, "elapsed_time": "0:32:18", "remaining_time": "8:38:36"} +{"current_steps": 225, "total_steps": 3752, "loss": 0.5947, "lr": 2.3829787234042553e-05, "epoch": 0.4197761194029851, "percentage": 6.0, "elapsed_time": "0:32:55", "remaining_time": "8:36:01"} +{"current_steps": 230, "total_steps": 3752, "loss": 0.6226, "lr": 2.4361702127659578e-05, "epoch": 0.4291044776119403, "percentage": 6.13, "elapsed_time": "0:33:36", "remaining_time": "8:34:46"} +{"current_steps": 235, "total_steps": 3752, "loss": 0.5873, "lr": 2.48936170212766e-05, "epoch": 0.43843283582089554, "percentage": 6.26, "elapsed_time": "0:34:13", "remaining_time": "8:32:12"} +{"current_steps": 240, "total_steps": 3752, "loss": 0.6136, "lr": 2.5425531914893616e-05, "epoch": 0.44776119402985076, "percentage": 6.4, "elapsed_time": "0:34:51", "remaining_time": "8:30:01"} +{"current_steps": 245, "total_steps": 3752, "loss": 0.6212, "lr": 2.595744680851064e-05, "epoch": 0.457089552238806, "percentage": 6.53, "elapsed_time": "0:35:24", "remaining_time": "8:26:57"} +{"current_steps": 250, "total_steps": 3752, "loss": 0.5631, "lr": 2.6489361702127664e-05, "epoch": 0.4664179104477612, "percentage": 6.66, "elapsed_time": "0:36:10", "remaining_time": "8:26:49"} +{"current_steps": 255, "total_steps": 3752, "loss": 0.5797, "lr": 2.702127659574468e-05, "epoch": 0.47574626865671643, "percentage": 6.8, "elapsed_time": "0:36:53", "remaining_time": "8:25:52"} +{"current_steps": 260, "total_steps": 3752, "loss": 0.5974, "lr": 2.7553191489361702e-05, "epoch": 0.48507462686567165, "percentage": 6.93, "elapsed_time": "0:37:40", "remaining_time": "8:26:01"} +{"current_steps": 265, "total_steps": 3752, "loss": 0.5887, "lr": 2.8085106382978727e-05, "epoch": 0.4944029850746269, "percentage": 7.06, "elapsed_time": "0:38:19", "remaining_time": "8:24:19"} +{"current_steps": 270, "total_steps": 3752, "loss": 0.6153, "lr": 2.8617021276595747e-05, "epoch": 0.503731343283582, "percentage": 7.2, "elapsed_time": "0:39:01", "remaining_time": "8:23:18"} +{"current_steps": 275, "total_steps": 3752, "loss": 0.6124, "lr": 2.9148936170212768e-05, "epoch": 0.5130597014925373, "percentage": 7.33, "elapsed_time": "0:39:42", "remaining_time": "8:22:04"} +{"current_steps": 280, "total_steps": 3752, "loss": 0.6264, "lr": 2.968085106382979e-05, "epoch": 0.5223880597014925, "percentage": 7.46, "elapsed_time": "0:40:14", "remaining_time": "8:19:04"} +{"current_steps": 285, "total_steps": 3752, "loss": 0.6174, "lr": 3.0212765957446813e-05, "epoch": 0.5317164179104478, "percentage": 7.6, "elapsed_time": "0:40:56", "remaining_time": "8:18:00"} +{"current_steps": 290, "total_steps": 3752, "loss": 0.6031, "lr": 3.074468085106383e-05, "epoch": 0.5410447761194029, "percentage": 7.73, "elapsed_time": "0:41:36", "remaining_time": "8:16:42"} +{"current_steps": 295, "total_steps": 3752, "loss": 0.5742, "lr": 3.127659574468085e-05, "epoch": 0.5503731343283582, "percentage": 7.86, "elapsed_time": "0:42:22", "remaining_time": "8:16:36"} +{"current_steps": 300, "total_steps": 3752, "loss": 0.6248, "lr": 3.180851063829788e-05, "epoch": 0.5597014925373134, "percentage": 8.0, "elapsed_time": "0:43:13", "remaining_time": "8:17:22"} +{"current_steps": 305, "total_steps": 3752, "loss": 0.5664, "lr": 3.234042553191489e-05, "epoch": 0.5690298507462687, "percentage": 8.13, "elapsed_time": "0:43:51", "remaining_time": "8:15:39"} +{"current_steps": 310, "total_steps": 3752, "loss": 0.6137, "lr": 3.2872340425531914e-05, "epoch": 0.5783582089552238, "percentage": 8.26, "elapsed_time": "0:44:44", "remaining_time": "8:16:44"} +{"current_steps": 315, "total_steps": 3752, "loss": 0.6167, "lr": 3.340425531914894e-05, "epoch": 0.5876865671641791, "percentage": 8.4, "elapsed_time": "0:45:15", "remaining_time": "8:13:45"} +{"current_steps": 320, "total_steps": 3752, "loss": 0.6265, "lr": 3.393617021276596e-05, "epoch": 0.5970149253731343, "percentage": 8.53, "elapsed_time": "0:46:00", "remaining_time": "8:13:21"} +{"current_steps": 325, "total_steps": 3752, "loss": 0.5727, "lr": 3.446808510638298e-05, "epoch": 0.6063432835820896, "percentage": 8.66, "elapsed_time": "0:46:47", "remaining_time": "8:13:21"} +{"current_steps": 330, "total_steps": 3752, "loss": 0.6388, "lr": 3.5000000000000004e-05, "epoch": 0.6156716417910447, "percentage": 8.8, "elapsed_time": "0:47:35", "remaining_time": "8:13:34"} +{"current_steps": 335, "total_steps": 3752, "loss": 0.5431, "lr": 3.5531914893617025e-05, "epoch": 0.625, "percentage": 8.93, "elapsed_time": "0:48:08", "remaining_time": "8:11:00"} +{"current_steps": 340, "total_steps": 3752, "loss": 0.5734, "lr": 3.6063829787234045e-05, "epoch": 0.6343283582089553, "percentage": 9.06, "elapsed_time": "0:48:53", "remaining_time": "8:10:41"} +{"current_steps": 345, "total_steps": 3752, "loss": 0.5895, "lr": 3.6595744680851066e-05, "epoch": 0.6436567164179104, "percentage": 9.2, "elapsed_time": "0:49:30", "remaining_time": "8:08:53"} +{"current_steps": 350, "total_steps": 3752, "loss": 0.5856, "lr": 3.712765957446809e-05, "epoch": 0.6529850746268657, "percentage": 9.33, "elapsed_time": "0:50:11", "remaining_time": "8:07:56"} +{"current_steps": 355, "total_steps": 3752, "loss": 0.5908, "lr": 3.7659574468085114e-05, "epoch": 0.6623134328358209, "percentage": 9.46, "elapsed_time": "0:51:00", "remaining_time": "8:08:02"} +{"current_steps": 360, "total_steps": 3752, "loss": 0.6502, "lr": 3.819148936170213e-05, "epoch": 0.6716417910447762, "percentage": 9.59, "elapsed_time": "0:51:29", "remaining_time": "8:05:06"} +{"current_steps": 365, "total_steps": 3752, "loss": 0.5736, "lr": 3.872340425531915e-05, "epoch": 0.6809701492537313, "percentage": 9.73, "elapsed_time": "0:52:05", "remaining_time": "8:03:18"} +{"current_steps": 370, "total_steps": 3752, "loss": 0.5662, "lr": 3.925531914893618e-05, "epoch": 0.6902985074626866, "percentage": 9.86, "elapsed_time": "0:52:50", "remaining_time": "8:03:00"} +{"current_steps": 375, "total_steps": 3752, "loss": 0.559, "lr": 3.978723404255319e-05, "epoch": 0.6996268656716418, "percentage": 9.99, "elapsed_time": "0:53:36", "remaining_time": "8:02:48"} +{"current_steps": 380, "total_steps": 3752, "loss": 0.5744, "lr": 3.999992206418221e-05, "epoch": 0.7089552238805971, "percentage": 10.13, "elapsed_time": "0:54:09", "remaining_time": "8:00:34"} +{"current_steps": 385, "total_steps": 3752, "loss": 0.611, "lr": 3.999944579193981e-05, "epoch": 0.7182835820895522, "percentage": 10.26, "elapsed_time": "0:54:45", "remaining_time": "7:58:55"} +{"current_steps": 390, "total_steps": 3752, "loss": 0.5903, "lr": 3.999853655542977e-05, "epoch": 0.7276119402985075, "percentage": 10.39, "elapsed_time": "0:55:25", "remaining_time": "7:57:51"} +{"current_steps": 395, "total_steps": 3752, "loss": 0.5938, "lr": 3.999719437433601e-05, "epoch": 0.7369402985074627, "percentage": 10.53, "elapsed_time": "0:56:09", "remaining_time": "7:57:12"} +{"current_steps": 400, "total_steps": 3752, "loss": 0.5653, "lr": 3.999541927771513e-05, "epoch": 0.746268656716418, "percentage": 10.66, "elapsed_time": "0:56:50", "remaining_time": "7:56:21"} +{"current_steps": 405, "total_steps": 3752, "loss": 0.5915, "lr": 3.999321130399588e-05, "epoch": 0.7555970149253731, "percentage": 10.79, "elapsed_time": "0:57:34", "remaining_time": "7:55:49"} +{"current_steps": 410, "total_steps": 3752, "loss": 0.6022, "lr": 3.9990570500978275e-05, "epoch": 0.7649253731343284, "percentage": 10.93, "elapsed_time": "0:58:13", "remaining_time": "7:54:34"} +{"current_steps": 415, "total_steps": 3752, "loss": 0.5879, "lr": 3.998749692583255e-05, "epoch": 0.7742537313432836, "percentage": 11.06, "elapsed_time": "0:59:04", "remaining_time": "7:55:02"} +{"current_steps": 420, "total_steps": 3752, "loss": 0.5516, "lr": 3.9983990645097977e-05, "epoch": 0.7835820895522388, "percentage": 11.19, "elapsed_time": "0:59:43", "remaining_time": "7:53:48"} +{"current_steps": 425, "total_steps": 3752, "loss": 0.5906, "lr": 3.9980051734681365e-05, "epoch": 0.792910447761194, "percentage": 11.33, "elapsed_time": "1:00:27", "remaining_time": "7:53:13"} +{"current_steps": 430, "total_steps": 3752, "loss": 0.571, "lr": 3.9975680279855453e-05, "epoch": 0.8022388059701493, "percentage": 11.46, "elapsed_time": "1:01:07", "remaining_time": "7:52:16"} +{"current_steps": 435, "total_steps": 3752, "loss": 0.5607, "lr": 3.9970876375257044e-05, "epoch": 0.8115671641791045, "percentage": 11.59, "elapsed_time": "1:01:56", "remaining_time": "7:52:20"} +{"current_steps": 440, "total_steps": 3752, "loss": 0.58, "lr": 3.996564012488497e-05, "epoch": 0.8208955223880597, "percentage": 11.73, "elapsed_time": "1:02:49", "remaining_time": "7:52:53"} +{"current_steps": 445, "total_steps": 3752, "loss": 0.5774, "lr": 3.995997164209785e-05, "epoch": 0.8302238805970149, "percentage": 11.86, "elapsed_time": "1:03:30", "remaining_time": "7:51:59"} +{"current_steps": 450, "total_steps": 3752, "loss": 0.5685, "lr": 3.99538710496116e-05, "epoch": 0.8395522388059702, "percentage": 11.99, "elapsed_time": "1:04:07", "remaining_time": "7:50:33"} +{"current_steps": 455, "total_steps": 3752, "loss": 0.5592, "lr": 3.9947338479496813e-05, "epoch": 0.8488805970149254, "percentage": 12.13, "elapsed_time": "1:04:44", "remaining_time": "7:49:08"} +{"current_steps": 460, "total_steps": 3752, "loss": 0.5656, "lr": 3.994037407317588e-05, "epoch": 0.8582089552238806, "percentage": 12.26, "elapsed_time": "1:05:21", "remaining_time": "7:47:44"} +{"current_steps": 465, "total_steps": 3752, "loss": 0.6036, "lr": 3.993297798141992e-05, "epoch": 0.8675373134328358, "percentage": 12.39, "elapsed_time": "1:06:02", "remaining_time": "7:46:52"} +{"current_steps": 470, "total_steps": 3752, "loss": 0.6171, "lr": 3.9925150364345556e-05, "epoch": 0.8768656716417911, "percentage": 12.53, "elapsed_time": "1:06:50", "remaining_time": "7:46:45"} +{"current_steps": 475, "total_steps": 3752, "loss": 0.5776, "lr": 3.991689139141141e-05, "epoch": 0.8861940298507462, "percentage": 12.66, "elapsed_time": "1:07:29", "remaining_time": "7:45:36"} +{"current_steps": 480, "total_steps": 3752, "loss": 0.5608, "lr": 3.9908201241414434e-05, "epoch": 0.8955223880597015, "percentage": 12.79, "elapsed_time": "1:08:16", "remaining_time": "7:45:21"} +{"current_steps": 485, "total_steps": 3752, "loss": 0.5632, "lr": 3.9899080102486065e-05, "epoch": 0.9048507462686567, "percentage": 12.93, "elapsed_time": "1:08:55", "remaining_time": "7:44:18"} +{"current_steps": 490, "total_steps": 3752, "loss": 0.5421, "lr": 3.988952817208815e-05, "epoch": 0.914179104477612, "percentage": 13.06, "elapsed_time": "1:09:49", "remaining_time": "7:44:52"} +{"current_steps": 495, "total_steps": 3752, "loss": 0.5646, "lr": 3.987954565700863e-05, "epoch": 0.9235074626865671, "percentage": 13.19, "elapsed_time": "1:10:40", "remaining_time": "7:45:02"} +{"current_steps": 500, "total_steps": 3752, "loss": 0.5626, "lr": 3.9869132773357115e-05, "epoch": 0.9328358208955224, "percentage": 13.33, "elapsed_time": "1:11:20", "remaining_time": "7:44:00"} +{"current_steps": 505, "total_steps": 3752, "loss": 0.5613, "lr": 3.9858289746560183e-05, "epoch": 0.9421641791044776, "percentage": 13.46, "elapsed_time": "1:12:03", "remaining_time": "7:43:21"} +{"current_steps": 510, "total_steps": 3752, "loss": 0.5653, "lr": 3.98470168113565e-05, "epoch": 0.9514925373134329, "percentage": 13.59, "elapsed_time": "1:12:50", "remaining_time": "7:43:04"} +{"current_steps": 515, "total_steps": 3752, "loss": 0.5751, "lr": 3.983531421179172e-05, "epoch": 0.960820895522388, "percentage": 13.73, "elapsed_time": "1:13:36", "remaining_time": "7:42:40"} +{"current_steps": 520, "total_steps": 3752, "loss": 0.5968, "lr": 3.9823182201213245e-05, "epoch": 0.9701492537313433, "percentage": 13.86, "elapsed_time": "1:14:16", "remaining_time": "7:41:38"} +{"current_steps": 525, "total_steps": 3752, "loss": 0.5553, "lr": 3.981062104226471e-05, "epoch": 0.9794776119402985, "percentage": 13.99, "elapsed_time": "1:14:59", "remaining_time": "7:40:59"} +{"current_steps": 530, "total_steps": 3752, "loss": 0.6386, "lr": 3.97976310068803e-05, "epoch": 0.9888059701492538, "percentage": 14.13, "elapsed_time": "1:15:52", "remaining_time": "7:41:15"} +{"current_steps": 535, "total_steps": 3752, "loss": 0.5233, "lr": 3.978421237627886e-05, "epoch": 0.9981343283582089, "percentage": 14.26, "elapsed_time": "1:16:37", "remaining_time": "7:40:43"} +{"current_steps": 540, "total_steps": 3752, "loss": 0.553, "lr": 3.977036544095781e-05, "epoch": 1.007462686567164, "percentage": 14.39, "elapsed_time": "1:17:32", "remaining_time": "7:41:11"} +{"current_steps": 545, "total_steps": 3752, "loss": 0.5425, "lr": 3.975609050068689e-05, "epoch": 1.0167910447761195, "percentage": 14.53, "elapsed_time": "1:18:15", "remaining_time": "7:40:27"} +{"current_steps": 550, "total_steps": 3752, "loss": 0.5363, "lr": 3.97413878645016e-05, "epoch": 1.0261194029850746, "percentage": 14.66, "elapsed_time": "1:19:03", "remaining_time": "7:40:17"} +{"current_steps": 555, "total_steps": 3752, "loss": 0.5389, "lr": 3.972625785069657e-05, "epoch": 1.0354477611940298, "percentage": 14.79, "elapsed_time": "1:19:48", "remaining_time": "7:39:43"} +{"current_steps": 560, "total_steps": 3752, "loss": 0.5043, "lr": 3.971070078681864e-05, "epoch": 1.044776119402985, "percentage": 14.93, "elapsed_time": "1:20:32", "remaining_time": "7:39:07"} +{"current_steps": 565, "total_steps": 3752, "loss": 0.537, "lr": 3.969471700965979e-05, "epoch": 1.0541044776119404, "percentage": 15.06, "elapsed_time": "1:21:21", "remaining_time": "7:38:56"} +{"current_steps": 570, "total_steps": 3752, "loss": 0.5388, "lr": 3.967830686524982e-05, "epoch": 1.0634328358208955, "percentage": 15.19, "elapsed_time": "1:22:11", "remaining_time": "7:38:50"} +{"current_steps": 575, "total_steps": 3752, "loss": 0.5375, "lr": 3.966147070884888e-05, "epoch": 1.0727611940298507, "percentage": 15.33, "elapsed_time": "1:23:00", "remaining_time": "7:38:38"} +{"current_steps": 580, "total_steps": 3752, "loss": 0.5429, "lr": 3.964420890493978e-05, "epoch": 1.0820895522388059, "percentage": 15.46, "elapsed_time": "1:23:52", "remaining_time": "7:38:40"} +{"current_steps": 585, "total_steps": 3752, "loss": 0.5274, "lr": 3.9626521827220096e-05, "epoch": 1.0914179104477613, "percentage": 15.59, "elapsed_time": "1:24:38", "remaining_time": "7:38:15"} +{"current_steps": 590, "total_steps": 3752, "loss": 0.5418, "lr": 3.960840985859406e-05, "epoch": 1.1007462686567164, "percentage": 15.72, "elapsed_time": "1:25:18", "remaining_time": "7:37:10"} +{"current_steps": 595, "total_steps": 3752, "loss": 0.5166, "lr": 3.95898733911643e-05, "epoch": 1.1100746268656716, "percentage": 15.86, "elapsed_time": "1:25:55", "remaining_time": "7:35:55"} +{"current_steps": 600, "total_steps": 3752, "loss": 0.5539, "lr": 3.957091282622335e-05, "epoch": 1.1194029850746268, "percentage": 15.99, "elapsed_time": "1:26:38", "remaining_time": "7:35:08"} +{"current_steps": 605, "total_steps": 3752, "loss": 0.5375, "lr": 3.955152857424493e-05, "epoch": 1.1287313432835822, "percentage": 16.12, "elapsed_time": "1:27:55", "remaining_time": "7:37:19"} +{"current_steps": 610, "total_steps": 3752, "loss": 0.5554, "lr": 3.953172105487509e-05, "epoch": 1.1380597014925373, "percentage": 16.26, "elapsed_time": "1:28:34", "remaining_time": "7:36:15"} +{"current_steps": 615, "total_steps": 3752, "loss": 0.5316, "lr": 3.951149069692312e-05, "epoch": 1.1473880597014925, "percentage": 16.39, "elapsed_time": "1:29:29", "remaining_time": "7:36:30"} +{"current_steps": 620, "total_steps": 3752, "loss": 0.5774, "lr": 3.9490837938352267e-05, "epoch": 1.1567164179104479, "percentage": 16.52, "elapsed_time": "1:30:17", "remaining_time": "7:36:05"} +{"current_steps": 625, "total_steps": 3752, "loss": 0.555, "lr": 3.946976322627024e-05, "epoch": 1.166044776119403, "percentage": 16.66, "elapsed_time": "1:30:51", "remaining_time": "7:34:33"} +{"current_steps": 630, "total_steps": 3752, "loss": 0.5606, "lr": 3.944826701691955e-05, "epoch": 1.1753731343283582, "percentage": 16.79, "elapsed_time": "1:31:26", "remaining_time": "7:33:09"} +{"current_steps": 635, "total_steps": 3752, "loss": 0.5389, "lr": 3.942634977566761e-05, "epoch": 1.1847014925373134, "percentage": 16.92, "elapsed_time": "1:32:04", "remaining_time": "7:31:58"} +{"current_steps": 640, "total_steps": 3752, "loss": 0.6084, "lr": 3.9404011976996696e-05, "epoch": 1.1940298507462686, "percentage": 17.06, "elapsed_time": "1:32:42", "remaining_time": "7:30:46"} +{"current_steps": 645, "total_steps": 3752, "loss": 0.5168, "lr": 3.938125410449364e-05, "epoch": 1.203358208955224, "percentage": 17.19, "elapsed_time": "1:33:27", "remaining_time": "7:30:11"} +{"current_steps": 650, "total_steps": 3752, "loss": 0.5465, "lr": 3.935807665083936e-05, "epoch": 1.212686567164179, "percentage": 17.32, "elapsed_time": "1:34:00", "remaining_time": "7:28:36"} +{"current_steps": 655, "total_steps": 3752, "loss": 0.5323, "lr": 3.933448011779824e-05, "epoch": 1.2220149253731343, "percentage": 17.46, "elapsed_time": "1:34:44", "remaining_time": "7:27:56"} +{"current_steps": 660, "total_steps": 3752, "loss": 0.5395, "lr": 3.93104650162072e-05, "epoch": 1.2313432835820897, "percentage": 17.59, "elapsed_time": "1:35:37", "remaining_time": "7:27:58"} +{"current_steps": 665, "total_steps": 3752, "loss": 0.5113, "lr": 3.928603186596471e-05, "epoch": 1.2406716417910448, "percentage": 17.72, "elapsed_time": "1:36:19", "remaining_time": "7:27:08"} +{"current_steps": 670, "total_steps": 3752, "loss": 0.5832, "lr": 3.926118119601946e-05, "epoch": 1.25, "percentage": 17.86, "elapsed_time": "1:37:04", "remaining_time": "7:26:33"} +{"current_steps": 675, "total_steps": 3752, "loss": 0.5735, "lr": 3.923591354435897e-05, "epoch": 1.2593283582089552, "percentage": 17.99, "elapsed_time": "1:37:47", "remaining_time": "7:25:48"} +{"current_steps": 680, "total_steps": 3752, "loss": 0.5395, "lr": 3.9210229457997916e-05, "epoch": 1.2686567164179103, "percentage": 18.12, "elapsed_time": "1:38:24", "remaining_time": "7:24:33"} +{"current_steps": 685, "total_steps": 3752, "loss": 0.4891, "lr": 3.9184129492966286e-05, "epoch": 1.2779850746268657, "percentage": 18.26, "elapsed_time": "1:39:07", "remaining_time": "7:23:47"} +{"current_steps": 690, "total_steps": 3752, "loss": 0.522, "lr": 3.915761421429735e-05, "epoch": 1.287313432835821, "percentage": 18.39, "elapsed_time": "1:40:02", "remaining_time": "7:23:56"} +{"current_steps": 695, "total_steps": 3752, "loss": 0.5745, "lr": 3.913068419601542e-05, "epoch": 1.296641791044776, "percentage": 18.52, "elapsed_time": "1:40:38", "remaining_time": "7:22:40"} +{"current_steps": 700, "total_steps": 3752, "loss": 0.5773, "lr": 3.910334002112344e-05, "epoch": 1.3059701492537314, "percentage": 18.66, "elapsed_time": "1:41:20", "remaining_time": "7:21:50"} +{"current_steps": 705, "total_steps": 3752, "loss": 0.5229, "lr": 3.907558228159034e-05, "epoch": 1.3152985074626866, "percentage": 18.79, "elapsed_time": "1:42:05", "remaining_time": "7:21:15"} +{"current_steps": 710, "total_steps": 3752, "loss": 0.5317, "lr": 3.904741157833825e-05, "epoch": 1.3246268656716418, "percentage": 18.92, "elapsed_time": "1:42:52", "remaining_time": "7:20:45"} +{"current_steps": 715, "total_steps": 3752, "loss": 0.5331, "lr": 3.901882852122945e-05, "epoch": 1.333955223880597, "percentage": 19.06, "elapsed_time": "1:43:42", "remaining_time": "7:20:28"} +{"current_steps": 720, "total_steps": 3752, "loss": 0.5126, "lr": 3.8989833729053226e-05, "epoch": 1.3432835820895521, "percentage": 19.19, "elapsed_time": "1:44:24", "remaining_time": "7:19:42"} +{"current_steps": 725, "total_steps": 3752, "loss": 0.5341, "lr": 3.89604278295124e-05, "epoch": 1.3526119402985075, "percentage": 19.32, "elapsed_time": "1:45:04", "remaining_time": "7:18:43"} +{"current_steps": 730, "total_steps": 3752, "loss": 0.5368, "lr": 3.893061145920981e-05, "epoch": 1.3619402985074627, "percentage": 19.46, "elapsed_time": "1:45:42", "remaining_time": "7:17:36"} +{"current_steps": 735, "total_steps": 3752, "loss": 0.5101, "lr": 3.890038526363449e-05, "epoch": 1.3712686567164178, "percentage": 19.59, "elapsed_time": "1:46:36", "remaining_time": "7:17:35"} +{"current_steps": 740, "total_steps": 3752, "loss": 0.4972, "lr": 3.88697498971477e-05, "epoch": 1.3805970149253732, "percentage": 19.72, "elapsed_time": "1:47:23", "remaining_time": "7:17:08"} +{"current_steps": 745, "total_steps": 3752, "loss": 0.4791, "lr": 3.883870602296878e-05, "epoch": 1.3899253731343284, "percentage": 19.86, "elapsed_time": "1:48:06", "remaining_time": "7:16:23"} +{"current_steps": 750, "total_steps": 3752, "loss": 0.5674, "lr": 3.880725431316074e-05, "epoch": 1.3992537313432836, "percentage": 19.99, "elapsed_time": "1:48:53", "remaining_time": "7:15:49"} +{"current_steps": 755, "total_steps": 3752, "loss": 0.5356, "lr": 3.87753954486158e-05, "epoch": 1.4085820895522387, "percentage": 20.12, "elapsed_time": "1:49:32", "remaining_time": "7:14:50"} +{"current_steps": 760, "total_steps": 3752, "loss": 0.5496, "lr": 3.874313011904056e-05, "epoch": 1.417910447761194, "percentage": 20.26, "elapsed_time": "1:50:17", "remaining_time": "7:14:10"} +{"current_steps": 765, "total_steps": 3752, "loss": 0.5187, "lr": 3.871045902294112e-05, "epoch": 1.4272388059701493, "percentage": 20.39, "elapsed_time": "1:50:50", "remaining_time": "7:12:47"} +{"current_steps": 770, "total_steps": 3752, "loss": 0.5382, "lr": 3.867738286760793e-05, "epoch": 1.4365671641791045, "percentage": 20.52, "elapsed_time": "1:51:35", "remaining_time": "7:12:09"} +{"current_steps": 775, "total_steps": 3752, "loss": 0.518, "lr": 3.86439023691005e-05, "epoch": 1.4458955223880596, "percentage": 20.66, "elapsed_time": "1:52:22", "remaining_time": "7:11:41"} +{"current_steps": 780, "total_steps": 3752, "loss": 0.5527, "lr": 3.86100182522319e-05, "epoch": 1.455223880597015, "percentage": 20.79, "elapsed_time": "1:53:01", "remaining_time": "7:10:38"} +{"current_steps": 785, "total_steps": 3752, "loss": 0.5226, "lr": 3.857573125055304e-05, "epoch": 1.4645522388059702, "percentage": 20.92, "elapsed_time": "1:53:45", "remaining_time": "7:09:56"} +{"current_steps": 790, "total_steps": 3752, "loss": 0.548, "lr": 3.8541042106336825e-05, "epoch": 1.4738805970149254, "percentage": 21.06, "elapsed_time": "1:54:24", "remaining_time": "7:08:58"} +{"current_steps": 795, "total_steps": 3752, "loss": 0.5544, "lr": 3.850595157056206e-05, "epoch": 1.4832089552238805, "percentage": 21.19, "elapsed_time": "1:55:00", "remaining_time": "7:07:46"} +{"current_steps": 800, "total_steps": 3752, "loss": 0.5091, "lr": 3.84704604028972e-05, "epoch": 1.4925373134328357, "percentage": 21.32, "elapsed_time": "1:55:56", "remaining_time": "7:07:48"} +{"current_steps": 805, "total_steps": 3752, "loss": 0.529, "lr": 3.84345693716839e-05, "epoch": 1.501865671641791, "percentage": 21.46, "elapsed_time": "1:56:33", "remaining_time": "7:06:41"} +{"current_steps": 810, "total_steps": 3752, "loss": 0.5368, "lr": 3.8398279253920403e-05, "epoch": 1.5111940298507462, "percentage": 21.59, "elapsed_time": "1:57:10", "remaining_time": "7:05:34"} +{"current_steps": 815, "total_steps": 3752, "loss": 0.535, "lr": 3.8361590835244696e-05, "epoch": 1.5205223880597014, "percentage": 21.72, "elapsed_time": "1:57:56", "remaining_time": "7:05:01"} +{"current_steps": 820, "total_steps": 3752, "loss": 0.533, "lr": 3.8324504909917486e-05, "epoch": 1.5298507462686568, "percentage": 21.86, "elapsed_time": "1:58:40", "remaining_time": "7:04:20"} +{"current_steps": 825, "total_steps": 3752, "loss": 0.5049, "lr": 3.8287022280805064e-05, "epoch": 1.539179104477612, "percentage": 21.99, "elapsed_time": "1:59:25", "remaining_time": "7:03:43"} +{"current_steps": 830, "total_steps": 3752, "loss": 0.5542, "lr": 3.824914375936186e-05, "epoch": 1.5485074626865671, "percentage": 22.12, "elapsed_time": "2:00:02", "remaining_time": "7:02:36"} +{"current_steps": 835, "total_steps": 3752, "loss": 0.5247, "lr": 3.8210870165612913e-05, "epoch": 1.5578358208955225, "percentage": 22.25, "elapsed_time": "2:00:45", "remaining_time": "7:01:51"} +{"current_steps": 840, "total_steps": 3752, "loss": 0.525, "lr": 3.817220232813611e-05, "epoch": 1.5671641791044775, "percentage": 22.39, "elapsed_time": "2:01:36", "remaining_time": "7:01:36"} +{"current_steps": 845, "total_steps": 3752, "loss": 0.5285, "lr": 3.8133141084044243e-05, "epoch": 1.5764925373134329, "percentage": 22.52, "elapsed_time": "2:02:07", "remaining_time": "7:00:09"} +{"current_steps": 850, "total_steps": 3752, "loss": 0.5223, "lr": 3.809368727896689e-05, "epoch": 1.585820895522388, "percentage": 22.65, "elapsed_time": "2:02:45", "remaining_time": "6:59:05"} +{"current_steps": 855, "total_steps": 3752, "loss": 0.5296, "lr": 3.805384176703211e-05, "epoch": 1.5951492537313432, "percentage": 22.79, "elapsed_time": "2:03:27", "remaining_time": "6:58:18"} +{"current_steps": 860, "total_steps": 3752, "loss": 0.5587, "lr": 3.8013605410847956e-05, "epoch": 1.6044776119402986, "percentage": 22.92, "elapsed_time": "2:04:21", "remaining_time": "6:58:11"} +{"current_steps": 865, "total_steps": 3752, "loss": 0.546, "lr": 3.797297908148379e-05, "epoch": 1.6138059701492538, "percentage": 23.05, "elapsed_time": "2:05:03", "remaining_time": "6:57:22"} +{"current_steps": 870, "total_steps": 3752, "loss": 0.5525, "lr": 3.793196365845142e-05, "epoch": 1.623134328358209, "percentage": 23.19, "elapsed_time": "2:05:45", "remaining_time": "6:56:35"} +{"current_steps": 875, "total_steps": 3752, "loss": 0.5657, "lr": 3.789056002968609e-05, "epoch": 1.6324626865671643, "percentage": 23.32, "elapsed_time": "2:06:22", "remaining_time": "6:55:31"} +{"current_steps": 880, "total_steps": 3752, "loss": 0.4998, "lr": 3.784876909152723e-05, "epoch": 1.6417910447761193, "percentage": 23.45, "elapsed_time": "2:07:00", "remaining_time": "6:54:30"} +{"current_steps": 885, "total_steps": 3752, "loss": 0.5349, "lr": 3.7806591748699056e-05, "epoch": 1.6511194029850746, "percentage": 23.59, "elapsed_time": "2:07:48", "remaining_time": "6:54:01"} +{"current_steps": 890, "total_steps": 3752, "loss": 0.5024, "lr": 3.776402891429098e-05, "epoch": 1.6604477611940298, "percentage": 23.72, "elapsed_time": "2:08:31", "remaining_time": "6:53:17"} +{"current_steps": 895, "total_steps": 3752, "loss": 0.5186, "lr": 3.772108150973786e-05, "epoch": 1.669776119402985, "percentage": 23.85, "elapsed_time": "2:09:12", "remaining_time": "6:52:25"} +{"current_steps": 900, "total_steps": 3752, "loss": 0.6051, "lr": 3.767775046480004e-05, "epoch": 1.6791044776119404, "percentage": 23.99, "elapsed_time": "2:09:56", "remaining_time": "6:51:46"} +{"current_steps": 905, "total_steps": 3752, "loss": 0.5367, "lr": 3.7634036717543224e-05, "epoch": 1.6884328358208955, "percentage": 24.12, "elapsed_time": "2:10:40", "remaining_time": "6:51:04"} +{"current_steps": 910, "total_steps": 3752, "loss": 0.5028, "lr": 3.7589941214318156e-05, "epoch": 1.6977611940298507, "percentage": 24.25, "elapsed_time": "2:11:20", "remaining_time": "6:50:12"} +{"current_steps": 915, "total_steps": 3752, "loss": 0.5176, "lr": 3.754546490974016e-05, "epoch": 1.707089552238806, "percentage": 24.39, "elapsed_time": "2:11:59", "remaining_time": "6:49:16"} +{"current_steps": 920, "total_steps": 3752, "loss": 0.5256, "lr": 3.750060876666844e-05, "epoch": 1.716417910447761, "percentage": 24.52, "elapsed_time": "2:12:35", "remaining_time": "6:48:08"} +{"current_steps": 925, "total_steps": 3752, "loss": 0.5176, "lr": 3.7455373756185276e-05, "epoch": 1.7257462686567164, "percentage": 24.65, "elapsed_time": "2:13:20", "remaining_time": "6:47:30"} +{"current_steps": 930, "total_steps": 3752, "loss": 0.508, "lr": 3.740976085757495e-05, "epoch": 1.7350746268656716, "percentage": 24.79, "elapsed_time": "2:14:04", "remaining_time": "6:46:51"} +{"current_steps": 935, "total_steps": 3752, "loss": 0.5459, "lr": 3.736377105830259e-05, "epoch": 1.7444029850746268, "percentage": 24.92, "elapsed_time": "2:14:38", "remaining_time": "6:45:40"} +{"current_steps": 940, "total_steps": 3752, "loss": 0.5267, "lr": 3.731740535399278e-05, "epoch": 1.7537313432835822, "percentage": 25.05, "elapsed_time": "2:15:20", "remaining_time": "6:44:52"} +{"current_steps": 945, "total_steps": 3752, "loss": 0.5194, "lr": 3.7270664748407985e-05, "epoch": 1.7630597014925373, "percentage": 25.19, "elapsed_time": "2:16:01", "remaining_time": "6:44:03"} +{"current_steps": 950, "total_steps": 3752, "loss": 0.526, "lr": 3.722355025342686e-05, "epoch": 1.7723880597014925, "percentage": 25.32, "elapsed_time": "2:16:51", "remaining_time": "6:43:39"} +{"current_steps": 955, "total_steps": 3752, "loss": 0.5453, "lr": 3.7176062889022296e-05, "epoch": 1.7817164179104479, "percentage": 25.45, "elapsed_time": "2:17:36", "remaining_time": "6:43:01"} +{"current_steps": 960, "total_steps": 3752, "loss": 0.5628, "lr": 3.7128203683239384e-05, "epoch": 1.7910447761194028, "percentage": 25.59, "elapsed_time": "2:18:20", "remaining_time": "6:42:20"} +{"current_steps": 965, "total_steps": 3752, "loss": 0.535, "lr": 3.7079973672173136e-05, "epoch": 1.8003731343283582, "percentage": 25.72, "elapsed_time": "2:18:53", "remaining_time": "6:41:08"} +{"current_steps": 970, "total_steps": 3752, "loss": 0.4916, "lr": 3.703137389994606e-05, "epoch": 1.8097014925373134, "percentage": 25.85, "elapsed_time": "2:19:47", "remaining_time": "6:40:54"} +{"current_steps": 975, "total_steps": 3752, "loss": 0.5622, "lr": 3.698240541868554e-05, "epoch": 1.8190298507462686, "percentage": 25.99, "elapsed_time": "2:20:26", "remaining_time": "6:40:00"} +{"current_steps": 980, "total_steps": 3752, "loss": 0.4906, "lr": 3.693306928850109e-05, "epoch": 1.828358208955224, "percentage": 26.12, "elapsed_time": "2:21:11", "remaining_time": "6:39:21"} +{"current_steps": 985, "total_steps": 3752, "loss": 0.5296, "lr": 3.688336657746137e-05, "epoch": 1.837686567164179, "percentage": 26.25, "elapsed_time": "2:21:56", "remaining_time": "6:38:43"} +{"current_steps": 990, "total_steps": 3752, "loss": 0.5396, "lr": 3.683329836157111e-05, "epoch": 1.8470149253731343, "percentage": 26.39, "elapsed_time": "2:22:45", "remaining_time": "6:38:18"} +{"current_steps": 995, "total_steps": 3752, "loss": 0.5106, "lr": 3.678286572474776e-05, "epoch": 1.8563432835820897, "percentage": 26.52, "elapsed_time": "2:23:17", "remaining_time": "6:37:01"} +{"current_steps": 1000, "total_steps": 3752, "loss": 0.5101, "lr": 3.673206975879804e-05, "epoch": 1.8656716417910446, "percentage": 26.65, "elapsed_time": "2:24:02", "remaining_time": "6:36:25"} +{"current_steps": 1005, "total_steps": 3752, "loss": 0.5369, "lr": 3.668091156339435e-05, "epoch": 1.875, "percentage": 26.79, "elapsed_time": "2:24:51", "remaining_time": "6:35:55"} +{"current_steps": 1010, "total_steps": 3752, "loss": 0.5012, "lr": 3.662939224605091e-05, "epoch": 1.8843283582089554, "percentage": 26.92, "elapsed_time": "2:25:45", "remaining_time": "6:35:43"} +{"current_steps": 1015, "total_steps": 3752, "loss": 0.5007, "lr": 3.657751292209979e-05, "epoch": 1.8936567164179103, "percentage": 27.05, "elapsed_time": "2:26:18", "remaining_time": "6:34:32"} +{"current_steps": 1020, "total_steps": 3752, "loss": 0.4933, "lr": 3.65252747146668e-05, "epoch": 1.9029850746268657, "percentage": 27.19, "elapsed_time": "2:27:03", "remaining_time": "6:33:52"} +{"current_steps": 1025, "total_steps": 3752, "loss": 0.5149, "lr": 3.647267875464714e-05, "epoch": 1.912313432835821, "percentage": 27.32, "elapsed_time": "2:27:45", "remaining_time": "6:33:06"} +{"current_steps": 1030, "total_steps": 3752, "loss": 0.52, "lr": 3.6419726180680914e-05, "epoch": 1.921641791044776, "percentage": 27.45, "elapsed_time": "2:28:27", "remaining_time": "6:32:19"} +{"current_steps": 1035, "total_steps": 3752, "loss": 0.5329, "lr": 3.636641813912851e-05, "epoch": 1.9309701492537314, "percentage": 27.59, "elapsed_time": "2:29:12", "remaining_time": "6:31:40"} +{"current_steps": 1040, "total_steps": 3752, "loss": 0.5328, "lr": 3.631275578404578e-05, "epoch": 1.9402985074626866, "percentage": 27.72, "elapsed_time": "2:29:55", "remaining_time": "6:30:57"} +{"current_steps": 1045, "total_steps": 3752, "loss": 0.5352, "lr": 3.6258740277159014e-05, "epoch": 1.9496268656716418, "percentage": 27.85, "elapsed_time": "2:30:40", "remaining_time": "6:30:20"} +{"current_steps": 1050, "total_steps": 3752, "loss": 0.5113, "lr": 3.6204372787839844e-05, "epoch": 1.9589552238805972, "percentage": 27.99, "elapsed_time": "2:31:22", "remaining_time": "6:29:32"} +{"current_steps": 1055, "total_steps": 3752, "loss": 0.564, "lr": 3.614965449307989e-05, "epoch": 1.9682835820895521, "percentage": 28.12, "elapsed_time": "2:31:59", "remaining_time": "6:28:33"} +{"current_steps": 1060, "total_steps": 3752, "loss": 0.5365, "lr": 3.609458657746531e-05, "epoch": 1.9776119402985075, "percentage": 28.25, "elapsed_time": "2:32:44", "remaining_time": "6:27:54"} +{"current_steps": 1065, "total_steps": 3752, "loss": 0.5223, "lr": 3.603917023315111e-05, "epoch": 1.9869402985074627, "percentage": 28.38, "elapsed_time": "2:33:23", "remaining_time": "6:27:01"} +{"current_steps": 1070, "total_steps": 3752, "loss": 0.4822, "lr": 3.598340665983539e-05, "epoch": 1.9962686567164178, "percentage": 28.52, "elapsed_time": "2:34:08", "remaining_time": "6:26:22"} +{"current_steps": 1075, "total_steps": 3752, "loss": 0.5024, "lr": 3.592729706473333e-05, "epoch": 2.0055970149253732, "percentage": 28.65, "elapsed_time": "2:34:52", "remaining_time": "6:25:40"} +{"current_steps": 1080, "total_steps": 3752, "loss": 0.5409, "lr": 3.587084266255108e-05, "epoch": 2.014925373134328, "percentage": 28.78, "elapsed_time": "2:35:29", "remaining_time": "6:24:41"} +{"current_steps": 1085, "total_steps": 3752, "loss": 0.4673, "lr": 3.581404467545946e-05, "epoch": 2.0242537313432836, "percentage": 28.92, "elapsed_time": "2:36:16", "remaining_time": "6:24:06"} +{"current_steps": 1090, "total_steps": 3752, "loss": 0.4907, "lr": 3.575690433306748e-05, "epoch": 2.033582089552239, "percentage": 29.05, "elapsed_time": "2:37:07", "remaining_time": "6:23:42"} +{"current_steps": 1095, "total_steps": 3752, "loss": 0.4737, "lr": 3.5699422872395735e-05, "epoch": 2.042910447761194, "percentage": 29.18, "elapsed_time": "2:37:47", "remaining_time": "6:22:53"} +{"current_steps": 1100, "total_steps": 3752, "loss": 0.5171, "lr": 3.564160153784965e-05, "epoch": 2.0522388059701493, "percentage": 29.32, "elapsed_time": "2:38:34", "remaining_time": "6:22:17"} +{"current_steps": 1105, "total_steps": 3752, "loss": 0.4906, "lr": 3.5583441581192484e-05, "epoch": 2.0615671641791047, "percentage": 29.45, "elapsed_time": "2:39:15", "remaining_time": "6:21:30"} +{"current_steps": 1110, "total_steps": 3752, "loss": 0.5119, "lr": 3.5524944261518295e-05, "epoch": 2.0708955223880596, "percentage": 29.58, "elapsed_time": "2:40:01", "remaining_time": "6:20:54"} +{"current_steps": 1115, "total_steps": 3752, "loss": 0.4977, "lr": 3.546611084522462e-05, "epoch": 2.080223880597015, "percentage": 29.72, "elapsed_time": "2:40:44", "remaining_time": "6:20:10"} +{"current_steps": 1120, "total_steps": 3752, "loss": 0.4752, "lr": 3.540694260598512e-05, "epoch": 2.08955223880597, "percentage": 29.85, "elapsed_time": "2:41:30", "remaining_time": "6:19:32"} +{"current_steps": 1125, "total_steps": 3752, "loss": 0.5074, "lr": 3.5347440824721946e-05, "epoch": 2.0988805970149254, "percentage": 29.98, "elapsed_time": "2:42:17", "remaining_time": "6:18:58"} +{"current_steps": 1130, "total_steps": 3752, "loss": 0.5059, "lr": 3.5287606789578046e-05, "epoch": 2.1082089552238807, "percentage": 30.12, "elapsed_time": "2:43:01", "remaining_time": "6:18:15"} +{"current_steps": 1135, "total_steps": 3752, "loss": 0.491, "lr": 3.522744179588928e-05, "epoch": 2.1175373134328357, "percentage": 30.25, "elapsed_time": "2:43:39", "remaining_time": "6:17:20"} +{"current_steps": 1140, "total_steps": 3752, "loss": 0.5014, "lr": 3.516694714615637e-05, "epoch": 2.126865671641791, "percentage": 30.38, "elapsed_time": "2:44:17", "remaining_time": "6:16:25"} +{"current_steps": 1145, "total_steps": 3752, "loss": 0.4995, "lr": 3.510612415001668e-05, "epoch": 2.1361940298507465, "percentage": 30.52, "elapsed_time": "2:45:09", "remaining_time": "6:16:02"} +{"current_steps": 1150, "total_steps": 3752, "loss": 0.4825, "lr": 3.50449741242159e-05, "epoch": 2.1455223880597014, "percentage": 30.65, "elapsed_time": "2:45:47", "remaining_time": "6:15:06"} +{"current_steps": 1155, "total_steps": 3752, "loss": 0.5049, "lr": 3.4983498392579526e-05, "epoch": 2.154850746268657, "percentage": 30.78, "elapsed_time": "2:46:30", "remaining_time": "6:14:23"} +{"current_steps": 1160, "total_steps": 3752, "loss": 0.5105, "lr": 3.492169828598419e-05, "epoch": 2.1641791044776117, "percentage": 30.92, "elapsed_time": "2:47:07", "remaining_time": "6:13:26"} +{"current_steps": 1165, "total_steps": 3752, "loss": 0.527, "lr": 3.485957514232886e-05, "epoch": 2.173507462686567, "percentage": 31.05, "elapsed_time": "2:48:00", "remaining_time": "6:13:04"} +{"current_steps": 1170, "total_steps": 3752, "loss": 0.529, "lr": 3.479713030650587e-05, "epoch": 2.1828358208955225, "percentage": 31.18, "elapsed_time": "2:48:44", "remaining_time": "6:12:22"} +{"current_steps": 1175, "total_steps": 3752, "loss": 0.4656, "lr": 3.4734365130371816e-05, "epoch": 2.1921641791044775, "percentage": 31.32, "elapsed_time": "2:49:24", "remaining_time": "6:11:32"} +{"current_steps": 1180, "total_steps": 3752, "loss": 0.4968, "lr": 3.467128097271828e-05, "epoch": 2.201492537313433, "percentage": 31.45, "elapsed_time": "2:50:02", "remaining_time": "6:10:37"} +{"current_steps": 1185, "total_steps": 3752, "loss": 0.5111, "lr": 3.46078791992424e-05, "epoch": 2.2108208955223883, "percentage": 31.58, "elapsed_time": "2:50:48", "remaining_time": "6:10:01"} +{"current_steps": 1190, "total_steps": 3752, "loss": 0.4717, "lr": 3.454416118251734e-05, "epoch": 2.220149253731343, "percentage": 31.72, "elapsed_time": "2:51:35", "remaining_time": "6:09:25"} +{"current_steps": 1195, "total_steps": 3752, "loss": 0.498, "lr": 3.448012830196255e-05, "epoch": 2.2294776119402986, "percentage": 31.85, "elapsed_time": "2:52:15", "remaining_time": "6:08:35"} +{"current_steps": 1200, "total_steps": 3752, "loss": 0.4906, "lr": 3.441578194381389e-05, "epoch": 2.2388059701492535, "percentage": 31.98, "elapsed_time": "2:52:52", "remaining_time": "6:07:38"} +{"current_steps": 1205, "total_steps": 3752, "loss": 0.4856, "lr": 3.435112350109367e-05, "epoch": 2.248134328358209, "percentage": 32.12, "elapsed_time": "2:54:10", "remaining_time": "6:08:10"} +{"current_steps": 1210, "total_steps": 3752, "loss": 0.4933, "lr": 3.428615437358043e-05, "epoch": 2.2574626865671643, "percentage": 32.25, "elapsed_time": "2:54:46", "remaining_time": "6:07:09"} +{"current_steps": 1215, "total_steps": 3752, "loss": 0.5027, "lr": 3.422087596777869e-05, "epoch": 2.2667910447761193, "percentage": 32.38, "elapsed_time": "2:55:27", "remaining_time": "6:06:22"} +{"current_steps": 1220, "total_steps": 3752, "loss": 0.4796, "lr": 3.415528969688846e-05, "epoch": 2.2761194029850746, "percentage": 32.52, "elapsed_time": "2:56:11", "remaining_time": "6:05:40"} +{"current_steps": 1225, "total_steps": 3752, "loss": 0.4555, "lr": 3.408939698077468e-05, "epoch": 2.28544776119403, "percentage": 32.65, "elapsed_time": "2:56:47", "remaining_time": "6:04:41"} +{"current_steps": 1230, "total_steps": 3752, "loss": 0.4917, "lr": 3.402319924593645e-05, "epoch": 2.294776119402985, "percentage": 32.78, "elapsed_time": "2:57:31", "remaining_time": "6:04:00"} +{"current_steps": 1235, "total_steps": 3752, "loss": 0.4669, "lr": 3.395669792547618e-05, "epoch": 2.3041044776119404, "percentage": 32.92, "elapsed_time": "2:58:26", "remaining_time": "6:03:40"} +{"current_steps": 1240, "total_steps": 3752, "loss": 0.4901, "lr": 3.388989445906853e-05, "epoch": 2.3134328358208958, "percentage": 33.05, "elapsed_time": "2:59:01", "remaining_time": "6:02:40"} +{"current_steps": 1245, "total_steps": 3752, "loss": 0.4864, "lr": 3.382279029292928e-05, "epoch": 2.3227611940298507, "percentage": 33.18, "elapsed_time": "2:59:49", "remaining_time": "6:02:06"} +{"current_steps": 1250, "total_steps": 3752, "loss": 0.4627, "lr": 3.3755386879783985e-05, "epoch": 2.332089552238806, "percentage": 33.32, "elapsed_time": "3:00:28", "remaining_time": "6:01:14"} +{"current_steps": 1255, "total_steps": 3752, "loss": 0.4742, "lr": 3.368768567883655e-05, "epoch": 2.341417910447761, "percentage": 33.45, "elapsed_time": "3:01:03", "remaining_time": "6:00:14"} +{"current_steps": 1260, "total_steps": 3752, "loss": 0.4907, "lr": 3.3619688155737646e-05, "epoch": 2.3507462686567164, "percentage": 33.58, "elapsed_time": "3:01:54", "remaining_time": "5:59:47"} +{"current_steps": 1265, "total_steps": 3752, "loss": 0.4711, "lr": 3.355139578255294e-05, "epoch": 2.360074626865672, "percentage": 33.72, "elapsed_time": "3:02:37", "remaining_time": "5:59:03"} +{"current_steps": 1270, "total_steps": 3752, "loss": 0.5177, "lr": 3.348281003773127e-05, "epoch": 2.3694029850746268, "percentage": 33.85, "elapsed_time": "3:03:20", "remaining_time": "5:58:18"} +{"current_steps": 1275, "total_steps": 3752, "loss": 0.5027, "lr": 3.3413932406072626e-05, "epoch": 2.378731343283582, "percentage": 33.98, "elapsed_time": "3:04:02", "remaining_time": "5:57:33"} +{"current_steps": 1280, "total_steps": 3752, "loss": 0.5029, "lr": 3.3344764378695984e-05, "epoch": 2.388059701492537, "percentage": 34.12, "elapsed_time": "3:04:45", "remaining_time": "5:56:47"} +{"current_steps": 1285, "total_steps": 3752, "loss": 0.4866, "lr": 3.3275307453007066e-05, "epoch": 2.3973880597014925, "percentage": 34.25, "elapsed_time": "3:05:25", "remaining_time": "5:55:58"} +{"current_steps": 1290, "total_steps": 3752, "loss": 0.5012, "lr": 3.3205563132665876e-05, "epoch": 2.406716417910448, "percentage": 34.38, "elapsed_time": "3:06:08", "remaining_time": "5:55:15"} +{"current_steps": 1295, "total_steps": 3752, "loss": 0.489, "lr": 3.313553292755419e-05, "epoch": 2.416044776119403, "percentage": 34.51, "elapsed_time": "3:06:48", "remaining_time": "5:54:24"} +{"current_steps": 1300, "total_steps": 3752, "loss": 0.4889, "lr": 3.306521835374284e-05, "epoch": 2.425373134328358, "percentage": 34.65, "elapsed_time": "3:07:30", "remaining_time": "5:53:40"} +{"current_steps": 1305, "total_steps": 3752, "loss": 0.4847, "lr": 3.2994620933458896e-05, "epoch": 2.4347014925373136, "percentage": 34.78, "elapsed_time": "3:08:15", "remaining_time": "5:53:00"} +{"current_steps": 1310, "total_steps": 3752, "loss": 0.4819, "lr": 3.292374219505274e-05, "epoch": 2.4440298507462686, "percentage": 34.91, "elapsed_time": "3:08:53", "remaining_time": "5:52:07"} +{"current_steps": 1315, "total_steps": 3752, "loss": 0.5311, "lr": 3.285258367296496e-05, "epoch": 2.453358208955224, "percentage": 35.05, "elapsed_time": "3:09:39", "remaining_time": "5:51:29"} +{"current_steps": 1320, "total_steps": 3752, "loss": 0.5187, "lr": 3.278114690769307e-05, "epoch": 2.4626865671641793, "percentage": 35.18, "elapsed_time": "3:10:24", "remaining_time": "5:50:48"} +{"current_steps": 1325, "total_steps": 3752, "loss": 0.4869, "lr": 3.2709433445758304e-05, "epoch": 2.4720149253731343, "percentage": 35.31, "elapsed_time": "3:10:57", "remaining_time": "5:49:46"} +{"current_steps": 1330, "total_steps": 3752, "loss": 0.4655, "lr": 3.263744483967198e-05, "epoch": 2.4813432835820897, "percentage": 35.45, "elapsed_time": "3:11:41", "remaining_time": "5:49:05"} +{"current_steps": 1335, "total_steps": 3752, "loss": 0.4912, "lr": 3.2565182647902015e-05, "epoch": 2.4906716417910446, "percentage": 35.58, "elapsed_time": "3:12:15", "remaining_time": "5:48:05"} +{"current_steps": 1340, "total_steps": 3752, "loss": 0.5031, "lr": 3.249264843483909e-05, "epoch": 2.5, "percentage": 35.71, "elapsed_time": "3:13:03", "remaining_time": "5:47:30"} +{"current_steps": 1345, "total_steps": 3752, "loss": 0.5416, "lr": 3.2419843770762836e-05, "epoch": 2.5093283582089554, "percentage": 35.85, "elapsed_time": "3:13:48", "remaining_time": "5:46:50"} +{"current_steps": 1350, "total_steps": 3752, "loss": 0.4774, "lr": 3.2346770231807843e-05, "epoch": 2.5186567164179103, "percentage": 35.98, "elapsed_time": "3:14:35", "remaining_time": "5:46:14"} +{"current_steps": 1355, "total_steps": 3752, "loss": 0.5156, "lr": 3.227342939992951e-05, "epoch": 2.5279850746268657, "percentage": 36.11, "elapsed_time": "3:15:16", "remaining_time": "5:45:25"} +{"current_steps": 1360, "total_steps": 3752, "loss": 0.4677, "lr": 3.2199822862869803e-05, "epoch": 2.5373134328358207, "percentage": 36.25, "elapsed_time": "3:16:02", "remaining_time": "5:44:48"} +{"current_steps": 1365, "total_steps": 3752, "loss": 0.4918, "lr": 3.212595221412291e-05, "epoch": 2.546641791044776, "percentage": 36.38, "elapsed_time": "3:16:46", "remaining_time": "5:44:06"} +{"current_steps": 1370, "total_steps": 3752, "loss": 0.4996, "lr": 3.2051819052900706e-05, "epoch": 2.5559701492537314, "percentage": 36.51, "elapsed_time": "3:17:33", "remaining_time": "5:43:29"} +{"current_steps": 1375, "total_steps": 3752, "loss": 0.4767, "lr": 3.1977424984098154e-05, "epoch": 2.5652985074626864, "percentage": 36.65, "elapsed_time": "3:18:13", "remaining_time": "5:42:40"} +{"current_steps": 1380, "total_steps": 3752, "loss": 0.501, "lr": 3.190277161825857e-05, "epoch": 2.574626865671642, "percentage": 36.78, "elapsed_time": "3:18:46", "remaining_time": "5:41:39"} +{"current_steps": 1385, "total_steps": 3752, "loss": 0.5063, "lr": 3.182786057153871e-05, "epoch": 2.583955223880597, "percentage": 36.91, "elapsed_time": "3:19:35", "remaining_time": "5:41:06"} +{"current_steps": 1390, "total_steps": 3752, "loss": 0.4559, "lr": 3.175269346567386e-05, "epoch": 2.593283582089552, "percentage": 37.05, "elapsed_time": "3:20:22", "remaining_time": "5:40:30"} +{"current_steps": 1395, "total_steps": 3752, "loss": 0.5141, "lr": 3.167727192794265e-05, "epoch": 2.6026119402985075, "percentage": 37.18, "elapsed_time": "3:21:04", "remaining_time": "5:39:43"} +{"current_steps": 1400, "total_steps": 3752, "loss": 0.4684, "lr": 3.1601597591131864e-05, "epoch": 2.611940298507463, "percentage": 37.31, "elapsed_time": "3:21:53", "remaining_time": "5:39:11"} +{"current_steps": 1405, "total_steps": 3752, "loss": 0.4915, "lr": 3.152567209350109e-05, "epoch": 2.621268656716418, "percentage": 37.45, "elapsed_time": "3:22:39", "remaining_time": "5:38:31"} +{"current_steps": 1410, "total_steps": 3752, "loss": 0.4866, "lr": 3.144949707874726e-05, "epoch": 2.6305970149253732, "percentage": 37.58, "elapsed_time": "3:23:21", "remaining_time": "5:37:47"} +{"current_steps": 1415, "total_steps": 3752, "loss": 0.4998, "lr": 3.137307419596904e-05, "epoch": 2.6399253731343286, "percentage": 37.71, "elapsed_time": "3:24:07", "remaining_time": "5:37:07"} +{"current_steps": 1420, "total_steps": 3752, "loss": 0.5035, "lr": 3.129640509963114e-05, "epoch": 2.6492537313432836, "percentage": 37.85, "elapsed_time": "3:24:51", "remaining_time": "5:36:26"} +{"current_steps": 1425, "total_steps": 3752, "loss": 0.4857, "lr": 3.121949144952853e-05, "epoch": 2.658582089552239, "percentage": 37.98, "elapsed_time": "3:25:34", "remaining_time": "5:35:42"} +{"current_steps": 1430, "total_steps": 3752, "loss": 0.5153, "lr": 3.1142334910750426e-05, "epoch": 2.667910447761194, "percentage": 38.11, "elapsed_time": "3:26:24", "remaining_time": "5:35:10"} +{"current_steps": 1435, "total_steps": 3752, "loss": 0.503, "lr": 3.1064937153644366e-05, "epoch": 2.6772388059701493, "percentage": 38.25, "elapsed_time": "3:27:09", "remaining_time": "5:34:28"} +{"current_steps": 1440, "total_steps": 3752, "loss": 0.5105, "lr": 3.098729985377992e-05, "epoch": 2.6865671641791042, "percentage": 38.38, "elapsed_time": "3:27:54", "remaining_time": "5:33:48"} +{"current_steps": 1445, "total_steps": 3752, "loss": 0.4605, "lr": 3.090942469191251e-05, "epoch": 2.6958955223880596, "percentage": 38.51, "elapsed_time": "3:28:37", "remaining_time": "5:33:04"} +{"current_steps": 1450, "total_steps": 3752, "loss": 0.493, "lr": 3.083131335394696e-05, "epoch": 2.705223880597015, "percentage": 38.65, "elapsed_time": "3:29:17", "remaining_time": "5:32:15"} +{"current_steps": 1455, "total_steps": 3752, "loss": 0.5031, "lr": 3.075296753090105e-05, "epoch": 2.71455223880597, "percentage": 38.78, "elapsed_time": "3:29:59", "remaining_time": "5:31:30"} +{"current_steps": 1460, "total_steps": 3752, "loss": 0.4703, "lr": 3.0674388918868876e-05, "epoch": 2.7238805970149254, "percentage": 38.91, "elapsed_time": "3:30:42", "remaining_time": "5:30:47"} +{"current_steps": 1465, "total_steps": 3752, "loss": 0.5103, "lr": 3.0595579218984124e-05, "epoch": 2.7332089552238807, "percentage": 39.05, "elapsed_time": "3:31:17", "remaining_time": "5:29:50"} +{"current_steps": 1470, "total_steps": 3752, "loss": 0.4848, "lr": 3.0516540137383287e-05, "epoch": 2.7425373134328357, "percentage": 39.18, "elapsed_time": "3:31:59", "remaining_time": "5:29:05"} +{"current_steps": 1475, "total_steps": 3752, "loss": 0.4981, "lr": 3.0437273385168677e-05, "epoch": 2.751865671641791, "percentage": 39.31, "elapsed_time": "3:32:40", "remaining_time": "5:28:18"} +{"current_steps": 1480, "total_steps": 3752, "loss": 0.4851, "lr": 3.0357780678371413e-05, "epoch": 2.7611940298507465, "percentage": 39.45, "elapsed_time": "3:33:18", "remaining_time": "5:27:27"} +{"current_steps": 1485, "total_steps": 3752, "loss": 0.4653, "lr": 3.027806373791426e-05, "epoch": 2.7705223880597014, "percentage": 39.58, "elapsed_time": "3:33:53", "remaining_time": "5:26:30"} +{"current_steps": 1490, "total_steps": 3752, "loss": 0.4661, "lr": 3.0198124289574378e-05, "epoch": 2.779850746268657, "percentage": 39.71, "elapsed_time": "3:34:30", "remaining_time": "5:25:39"} +{"current_steps": 1495, "total_steps": 3752, "loss": 0.5151, "lr": 3.011796406394596e-05, "epoch": 2.789179104477612, "percentage": 39.85, "elapsed_time": "3:35:12", "remaining_time": "5:24:54"} +{"current_steps": 1500, "total_steps": 3752, "loss": 0.4883, "lr": 3.0037584796402767e-05, "epoch": 2.798507462686567, "percentage": 39.98, "elapsed_time": "3:36:01", "remaining_time": "5:24:19"} +{"current_steps": 1505, "total_steps": 3752, "loss": 0.5217, "lr": 2.9956988227060536e-05, "epoch": 2.8078358208955225, "percentage": 40.11, "elapsed_time": "3:36:44", "remaining_time": "5:23:35"} +{"current_steps": 1510, "total_steps": 3752, "loss": 0.5065, "lr": 2.9876176100739368e-05, "epoch": 2.8171641791044775, "percentage": 40.25, "elapsed_time": "3:37:33", "remaining_time": "5:23:01"} +{"current_steps": 1515, "total_steps": 3752, "loss": 0.4586, "lr": 2.979515016692589e-05, "epoch": 2.826492537313433, "percentage": 40.38, "elapsed_time": "3:38:20", "remaining_time": "5:22:23"} +{"current_steps": 1520, "total_steps": 3752, "loss": 0.4819, "lr": 2.9713912179735404e-05, "epoch": 2.835820895522388, "percentage": 40.51, "elapsed_time": "3:38:56", "remaining_time": "5:21:30"} +{"current_steps": 1525, "total_steps": 3752, "loss": 0.4723, "lr": 2.9632463897873935e-05, "epoch": 2.845149253731343, "percentage": 40.64, "elapsed_time": "3:39:42", "remaining_time": "5:20:51"} +{"current_steps": 1530, "total_steps": 3752, "loss": 0.5032, "lr": 2.9550807084600134e-05, "epoch": 2.8544776119402986, "percentage": 40.78, "elapsed_time": "3:40:27", "remaining_time": "5:20:09"} +{"current_steps": 1535, "total_steps": 3752, "loss": 0.4886, "lr": 2.946894350768709e-05, "epoch": 2.8638059701492535, "percentage": 40.91, "elapsed_time": "3:41:15", "remaining_time": "5:19:34"} +{"current_steps": 1540, "total_steps": 3752, "loss": 0.49, "lr": 2.9386874939384103e-05, "epoch": 2.873134328358209, "percentage": 41.04, "elapsed_time": "3:41:55", "remaining_time": "5:18:45"} +{"current_steps": 1545, "total_steps": 3752, "loss": 0.4836, "lr": 2.9304603156378283e-05, "epoch": 2.8824626865671643, "percentage": 41.18, "elapsed_time": "3:42:41", "remaining_time": "5:18:06"} +{"current_steps": 1550, "total_steps": 3752, "loss": 0.5168, "lr": 2.9222129939756104e-05, "epoch": 2.8917910447761193, "percentage": 41.31, "elapsed_time": "3:43:21", "remaining_time": "5:17:18"} +{"current_steps": 1555, "total_steps": 3752, "loss": 0.4895, "lr": 2.913945707496482e-05, "epoch": 2.9011194029850746, "percentage": 41.44, "elapsed_time": "3:44:07", "remaining_time": "5:16:38"} +{"current_steps": 1560, "total_steps": 3752, "loss": 0.471, "lr": 2.9056586351773854e-05, "epoch": 2.91044776119403, "percentage": 41.58, "elapsed_time": "3:44:50", "remaining_time": "5:15:56"} +{"current_steps": 1565, "total_steps": 3752, "loss": 0.4799, "lr": 2.8973519564236024e-05, "epoch": 2.919776119402985, "percentage": 41.71, "elapsed_time": "3:45:32", "remaining_time": "5:15:11"} +{"current_steps": 1570, "total_steps": 3752, "loss": 0.4801, "lr": 2.8890258510648685e-05, "epoch": 2.9291044776119404, "percentage": 41.84, "elapsed_time": "3:46:17", "remaining_time": "5:14:29"} +{"current_steps": 1575, "total_steps": 3752, "loss": 0.4695, "lr": 2.8806804993514852e-05, "epoch": 2.9384328358208958, "percentage": 41.98, "elapsed_time": "3:47:10", "remaining_time": "5:14:00"} +{"current_steps": 1580, "total_steps": 3752, "loss": 0.478, "lr": 2.8723160819504133e-05, "epoch": 2.9477611940298507, "percentage": 42.11, "elapsed_time": "3:47:50", "remaining_time": "5:13:12"} +{"current_steps": 1585, "total_steps": 3752, "loss": 0.4783, "lr": 2.8639327799413635e-05, "epoch": 2.957089552238806, "percentage": 42.24, "elapsed_time": "3:48:36", "remaining_time": "5:12:33"} +{"current_steps": 1590, "total_steps": 3752, "loss": 0.4676, "lr": 2.8555307748128752e-05, "epoch": 2.966417910447761, "percentage": 42.38, "elapsed_time": "3:49:27", "remaining_time": "5:11:59"} +{"current_steps": 1595, "total_steps": 3752, "loss": 0.4881, "lr": 2.847110248458389e-05, "epoch": 2.9757462686567164, "percentage": 42.51, "elapsed_time": "3:50:05", "remaining_time": "5:11:09"} +{"current_steps": 1600, "total_steps": 3752, "loss": 0.4796, "lr": 2.838671383172307e-05, "epoch": 2.9850746268656714, "percentage": 42.64, "elapsed_time": "3:50:48", "remaining_time": "5:10:26"} +{"current_steps": 1605, "total_steps": 3752, "loss": 0.4917, "lr": 2.830214361646049e-05, "epoch": 2.9944029850746268, "percentage": 42.78, "elapsed_time": "3:51:27", "remaining_time": "5:09:37"} +{"current_steps": 1610, "total_steps": 3752, "loss": 0.4777, "lr": 2.8217393669640944e-05, "epoch": 3.003731343283582, "percentage": 42.91, "elapsed_time": "3:52:04", "remaining_time": "5:08:45"} +{"current_steps": 1615, "total_steps": 3752, "loss": 0.4325, "lr": 2.8132465826000207e-05, "epoch": 3.013059701492537, "percentage": 43.04, "elapsed_time": "3:52:48", "remaining_time": "5:08:03"} +{"current_steps": 1620, "total_steps": 3752, "loss": 0.456, "lr": 2.8047361924125307e-05, "epoch": 3.0223880597014925, "percentage": 43.18, "elapsed_time": "3:53:33", "remaining_time": "5:07:23"} +{"current_steps": 1625, "total_steps": 3752, "loss": 0.4691, "lr": 2.7962083806414727e-05, "epoch": 3.031716417910448, "percentage": 43.31, "elapsed_time": "3:54:17", "remaining_time": "5:06:40"} +{"current_steps": 1630, "total_steps": 3752, "loss": 0.4593, "lr": 2.7876633319038514e-05, "epoch": 3.041044776119403, "percentage": 43.44, "elapsed_time": "3:54:51", "remaining_time": "5:05:44"} +{"current_steps": 1635, "total_steps": 3752, "loss": 0.4487, "lr": 2.779101231189831e-05, "epoch": 3.050373134328358, "percentage": 43.58, "elapsed_time": "3:55:33", "remaining_time": "5:05:00"} +{"current_steps": 1640, "total_steps": 3752, "loss": 0.4865, "lr": 2.7705222638587308e-05, "epoch": 3.0597014925373136, "percentage": 43.71, "elapsed_time": "3:56:15", "remaining_time": "5:04:14"} +{"current_steps": 1645, "total_steps": 3752, "loss": 0.4582, "lr": 2.761926615635014e-05, "epoch": 3.0690298507462686, "percentage": 43.84, "elapsed_time": "3:57:01", "remaining_time": "5:03:36"} +{"current_steps": 1650, "total_steps": 3752, "loss": 0.4923, "lr": 2.7533144726042632e-05, "epoch": 3.078358208955224, "percentage": 43.98, "elapsed_time": "3:57:31", "remaining_time": "5:02:35"} +{"current_steps": 1655, "total_steps": 3752, "loss": 0.4886, "lr": 2.7446860212091548e-05, "epoch": 3.0876865671641793, "percentage": 44.11, "elapsed_time": "3:58:08", "remaining_time": "5:01:44"} +{"current_steps": 1660, "total_steps": 3752, "loss": 0.4647, "lr": 2.736041448245422e-05, "epoch": 3.0970149253731343, "percentage": 44.24, "elapsed_time": "3:58:52", "remaining_time": "5:01:02"} +{"current_steps": 1665, "total_steps": 3752, "loss": 0.4409, "lr": 2.7273809408578107e-05, "epoch": 3.1063432835820897, "percentage": 44.38, "elapsed_time": "3:59:36", "remaining_time": "5:00:20"} +{"current_steps": 1670, "total_steps": 3752, "loss": 0.4832, "lr": 2.7187046865360285e-05, "epoch": 3.1156716417910446, "percentage": 44.51, "elapsed_time": "4:00:15", "remaining_time": "4:59:31"} +{"current_steps": 1675, "total_steps": 3752, "loss": 0.4952, "lr": 2.7100128731106847e-05, "epoch": 3.125, "percentage": 44.64, "elapsed_time": "4:00:57", "remaining_time": "4:58:46"} +{"current_steps": 1680, "total_steps": 3752, "loss": 0.4623, "lr": 2.7013056887492255e-05, "epoch": 3.1343283582089554, "percentage": 44.78, "elapsed_time": "4:01:39", "remaining_time": "4:58:02"} +{"current_steps": 1685, "total_steps": 3752, "loss": 0.4574, "lr": 2.6925833219518596e-05, "epoch": 3.1436567164179103, "percentage": 44.91, "elapsed_time": "4:02:16", "remaining_time": "4:57:11"} +{"current_steps": 1690, "total_steps": 3752, "loss": 0.4685, "lr": 2.6838459615474758e-05, "epoch": 3.1529850746268657, "percentage": 45.04, "elapsed_time": "4:02:53", "remaining_time": "4:56:21"} +{"current_steps": 1695, "total_steps": 3752, "loss": 0.4484, "lr": 2.6750937966895603e-05, "epoch": 3.1623134328358207, "percentage": 45.18, "elapsed_time": "4:03:37", "remaining_time": "4:55:39"} +{"current_steps": 1700, "total_steps": 3752, "loss": 0.4592, "lr": 2.666327016852094e-05, "epoch": 3.171641791044776, "percentage": 45.31, "elapsed_time": "4:04:19", "remaining_time": "4:54:55"} +{"current_steps": 1705, "total_steps": 3752, "loss": 0.471, "lr": 2.6575458118254576e-05, "epoch": 3.1809701492537314, "percentage": 45.44, "elapsed_time": "4:05:07", "remaining_time": "4:54:17"} +{"current_steps": 1710, "total_steps": 3752, "loss": 0.4889, "lr": 2.648750371712319e-05, "epoch": 3.1902985074626864, "percentage": 45.58, "elapsed_time": "4:05:48", "remaining_time": "4:53:32"} +{"current_steps": 1715, "total_steps": 3752, "loss": 0.4492, "lr": 2.6399408869235187e-05, "epoch": 3.199626865671642, "percentage": 45.71, "elapsed_time": "4:06:21", "remaining_time": "4:52:37"} +{"current_steps": 1720, "total_steps": 3752, "loss": 0.4636, "lr": 2.6311175481739473e-05, "epoch": 3.208955223880597, "percentage": 45.84, "elapsed_time": "4:07:08", "remaining_time": "4:51:58"} +{"current_steps": 1725, "total_steps": 3752, "loss": 0.4925, "lr": 2.6222805464784196e-05, "epoch": 3.218283582089552, "percentage": 45.98, "elapsed_time": "4:07:52", "remaining_time": "4:51:15"} +{"current_steps": 1730, "total_steps": 3752, "loss": 0.4569, "lr": 2.6134300731475337e-05, "epoch": 3.2276119402985075, "percentage": 46.11, "elapsed_time": "4:08:36", "remaining_time": "4:50:34"} +{"current_steps": 1735, "total_steps": 3752, "loss": 0.4482, "lr": 2.6045663197835352e-05, "epoch": 3.236940298507463, "percentage": 46.24, "elapsed_time": "4:09:15", "remaining_time": "4:49:45"} +{"current_steps": 1740, "total_steps": 3752, "loss": 0.4679, "lr": 2.5956894782761638e-05, "epoch": 3.246268656716418, "percentage": 46.38, "elapsed_time": "4:10:04", "remaining_time": "4:49:09"} +{"current_steps": 1745, "total_steps": 3752, "loss": 0.4582, "lr": 2.5867997407985063e-05, "epoch": 3.2555970149253732, "percentage": 46.51, "elapsed_time": "4:10:44", "remaining_time": "4:48:23"} +{"current_steps": 1750, "total_steps": 3752, "loss": 0.4797, "lr": 2.577897299802828e-05, "epoch": 3.264925373134328, "percentage": 46.64, "elapsed_time": "4:11:19", "remaining_time": "4:47:31"} +{"current_steps": 1755, "total_steps": 3752, "loss": 0.4729, "lr": 2.568982348016412e-05, "epoch": 3.2742537313432836, "percentage": 46.78, "elapsed_time": "4:12:04", "remaining_time": "4:46:50"} +{"current_steps": 1760, "total_steps": 3752, "loss": 0.4748, "lr": 2.560055078437385e-05, "epoch": 3.283582089552239, "percentage": 46.91, "elapsed_time": "4:12:43", "remaining_time": "4:46:02"} +{"current_steps": 1765, "total_steps": 3752, "loss": 0.4742, "lr": 2.5511156843305393e-05, "epoch": 3.292910447761194, "percentage": 47.04, "elapsed_time": "4:13:31", "remaining_time": "4:45:25"} +{"current_steps": 1770, "total_steps": 3752, "loss": 0.4334, "lr": 2.542164359223148e-05, "epoch": 3.3022388059701493, "percentage": 47.17, "elapsed_time": "4:14:18", "remaining_time": "4:44:46"} +{"current_steps": 1775, "total_steps": 3752, "loss": 0.46, "lr": 2.5332012969007786e-05, "epoch": 3.3115671641791042, "percentage": 47.31, "elapsed_time": "4:14:56", "remaining_time": "4:43:57"} +{"current_steps": 1780, "total_steps": 3752, "loss": 0.4698, "lr": 2.5242266914030916e-05, "epoch": 3.3208955223880596, "percentage": 47.44, "elapsed_time": "4:15:37", "remaining_time": "4:43:12"} +{"current_steps": 1785, "total_steps": 3752, "loss": 0.4746, "lr": 2.5152407370196467e-05, "epoch": 3.330223880597015, "percentage": 47.57, "elapsed_time": "4:16:28", "remaining_time": "4:42:37"} +{"current_steps": 1790, "total_steps": 3752, "loss": 0.4137, "lr": 2.5062436282856926e-05, "epoch": 3.33955223880597, "percentage": 47.71, "elapsed_time": "4:17:18", "remaining_time": "4:42:01"} +{"current_steps": 1795, "total_steps": 3752, "loss": 0.4259, "lr": 2.4972355599779566e-05, "epoch": 3.3488805970149254, "percentage": 47.84, "elapsed_time": "4:18:06", "remaining_time": "4:41:24"} +{"current_steps": 1800, "total_steps": 3752, "loss": 0.4675, "lr": 2.4882167271104272e-05, "epoch": 3.3582089552238807, "percentage": 47.97, "elapsed_time": "4:18:50", "remaining_time": "4:40:42"} +{"current_steps": 1805, "total_steps": 3752, "loss": 0.481, "lr": 2.4791873249301327e-05, "epoch": 3.3675373134328357, "percentage": 48.11, "elapsed_time": "4:20:18", "remaining_time": "4:40:47"} +{"current_steps": 1810, "total_steps": 3752, "loss": 0.4598, "lr": 2.4701475489129152e-05, "epoch": 3.376865671641791, "percentage": 48.24, "elapsed_time": "4:20:56", "remaining_time": "4:39:58"} +{"current_steps": 1815, "total_steps": 3752, "loss": 0.4507, "lr": 2.4610975947591988e-05, "epoch": 3.3861940298507465, "percentage": 48.37, "elapsed_time": "4:21:40", "remaining_time": "4:39:15"} +{"current_steps": 1820, "total_steps": 3752, "loss": 0.4549, "lr": 2.452037658389749e-05, "epoch": 3.3955223880597014, "percentage": 48.51, "elapsed_time": "4:22:18", "remaining_time": "4:38:27"} +{"current_steps": 1825, "total_steps": 3752, "loss": 0.4769, "lr": 2.4429679359414386e-05, "epoch": 3.404850746268657, "percentage": 48.64, "elapsed_time": "4:23:01", "remaining_time": "4:37:43"} +{"current_steps": 1830, "total_steps": 3752, "loss": 0.4503, "lr": 2.433888623762994e-05, "epoch": 3.4141791044776117, "percentage": 48.77, "elapsed_time": "4:23:46", "remaining_time": "4:37:01"} +{"current_steps": 1835, "total_steps": 3752, "loss": 0.4556, "lr": 2.42479991841075e-05, "epoch": 3.423507462686567, "percentage": 48.91, "elapsed_time": "4:24:36", "remaining_time": "4:36:25"} +{"current_steps": 1840, "total_steps": 3752, "loss": 0.4623, "lr": 2.4157020166443913e-05, "epoch": 3.4328358208955225, "percentage": 49.04, "elapsed_time": "4:25:16", "remaining_time": "4:35:39"} +{"current_steps": 1845, "total_steps": 3752, "loss": 0.4702, "lr": 2.4065951154226948e-05, "epoch": 3.4421641791044775, "percentage": 49.17, "elapsed_time": "4:26:00", "remaining_time": "4:34:56"} +{"current_steps": 1850, "total_steps": 3752, "loss": 0.4714, "lr": 2.3974794118992652e-05, "epoch": 3.451492537313433, "percentage": 49.31, "elapsed_time": "4:26:38", "remaining_time": "4:34:08"} +{"current_steps": 1855, "total_steps": 3752, "loss": 0.4685, "lr": 2.3883551034182647e-05, "epoch": 3.4608208955223883, "percentage": 49.44, "elapsed_time": "4:27:29", "remaining_time": "4:33:32"} +{"current_steps": 1860, "total_steps": 3752, "loss": 0.4578, "lr": 2.3792223875101446e-05, "epoch": 3.470149253731343, "percentage": 49.57, "elapsed_time": "4:28:05", "remaining_time": "4:32:41"} +{"current_steps": 1865, "total_steps": 3752, "loss": 0.4533, "lr": 2.3700814618873662e-05, "epoch": 3.4794776119402986, "percentage": 49.71, "elapsed_time": "4:28:45", "remaining_time": "4:31:55"} +{"current_steps": 1870, "total_steps": 3752, "loss": 0.4763, "lr": 2.3609325244401212e-05, "epoch": 3.4888059701492535, "percentage": 49.84, "elapsed_time": "4:29:27", "remaining_time": "4:31:11"} +{"current_steps": 1875, "total_steps": 3752, "loss": 0.4462, "lr": 2.3517757732320484e-05, "epoch": 3.498134328358209, "percentage": 49.97, "elapsed_time": "4:30:20", "remaining_time": "4:30:37"} +{"current_steps": 1880, "total_steps": 3752, "loss": 0.4811, "lr": 2.342611406495946e-05, "epoch": 3.5074626865671643, "percentage": 50.11, "elapsed_time": "4:30:55", "remaining_time": "4:29:46"} +{"current_steps": 1885, "total_steps": 3752, "loss": 0.453, "lr": 2.333439622629476e-05, "epoch": 3.5167910447761193, "percentage": 50.24, "elapsed_time": "4:31:29", "remaining_time": "4:28:53"} +{"current_steps": 1890, "total_steps": 3752, "loss": 0.4737, "lr": 2.3242606201908758e-05, "epoch": 3.5261194029850746, "percentage": 50.37, "elapsed_time": "4:32:16", "remaining_time": "4:28:14"} +{"current_steps": 1895, "total_steps": 3752, "loss": 0.4519, "lr": 2.3150745978946563e-05, "epoch": 3.53544776119403, "percentage": 50.51, "elapsed_time": "4:32:56", "remaining_time": "4:27:28"} +{"current_steps": 1900, "total_steps": 3752, "loss": 0.4539, "lr": 2.3058817546072976e-05, "epoch": 3.544776119402985, "percentage": 50.64, "elapsed_time": "4:33:40", "remaining_time": "4:26:46"} +{"current_steps": 1905, "total_steps": 3752, "loss": 0.4681, "lr": 2.2966822893429487e-05, "epoch": 3.5541044776119404, "percentage": 50.77, "elapsed_time": "4:34:14", "remaining_time": "4:25:53"} +{"current_steps": 1910, "total_steps": 3752, "loss": 0.4667, "lr": 2.287476401259114e-05, "epoch": 3.5634328358208958, "percentage": 50.91, "elapsed_time": "4:35:01", "remaining_time": "4:25:14"} +{"current_steps": 1915, "total_steps": 3752, "loss": 0.4426, "lr": 2.278264289652347e-05, "epoch": 3.5727611940298507, "percentage": 51.04, "elapsed_time": "4:35:49", "remaining_time": "4:24:35"} +{"current_steps": 1920, "total_steps": 3752, "loss": 0.4701, "lr": 2.269046153953931e-05, "epoch": 3.582089552238806, "percentage": 51.17, "elapsed_time": "4:36:40", "remaining_time": "4:23:59"} +{"current_steps": 1925, "total_steps": 3752, "loss": 0.478, "lr": 2.2598221937255645e-05, "epoch": 3.591417910447761, "percentage": 51.31, "elapsed_time": "4:37:23", "remaining_time": "4:23:16"} +{"current_steps": 1930, "total_steps": 3752, "loss": 0.4608, "lr": 2.2505926086550404e-05, "epoch": 3.6007462686567164, "percentage": 51.44, "elapsed_time": "4:38:05", "remaining_time": "4:22:31"} +{"current_steps": 1935, "total_steps": 3752, "loss": 0.4507, "lr": 2.2413575985519226e-05, "epoch": 3.6100746268656714, "percentage": 51.57, "elapsed_time": "4:38:49", "remaining_time": "4:21:49"} +{"current_steps": 1940, "total_steps": 3752, "loss": 0.4686, "lr": 2.2321173633432202e-05, "epoch": 3.6194029850746268, "percentage": 51.71, "elapsed_time": "4:39:33", "remaining_time": "4:21:06"} +{"current_steps": 1945, "total_steps": 3752, "loss": 0.4314, "lr": 2.2228721030690603e-05, "epoch": 3.628731343283582, "percentage": 51.84, "elapsed_time": "4:40:12", "remaining_time": "4:20:19"} +{"current_steps": 1950, "total_steps": 3752, "loss": 0.5067, "lr": 2.2136220178783557e-05, "epoch": 3.638059701492537, "percentage": 51.97, "elapsed_time": "4:40:56", "remaining_time": "4:19:37"} +{"current_steps": 1955, "total_steps": 3752, "loss": 0.4273, "lr": 2.2043673080244744e-05, "epoch": 3.6473880597014925, "percentage": 52.11, "elapsed_time": "4:41:44", "remaining_time": "4:18:57"} +{"current_steps": 1960, "total_steps": 3752, "loss": 0.483, "lr": 2.1951081738609012e-05, "epoch": 3.656716417910448, "percentage": 52.24, "elapsed_time": "4:42:22", "remaining_time": "4:18:10"} +{"current_steps": 1965, "total_steps": 3752, "loss": 0.4599, "lr": 2.1858448158369027e-05, "epoch": 3.666044776119403, "percentage": 52.37, "elapsed_time": "4:43:05", "remaining_time": "4:17:26"} +{"current_steps": 1970, "total_steps": 3752, "loss": 0.4606, "lr": 2.1765774344931877e-05, "epoch": 3.675373134328358, "percentage": 52.51, "elapsed_time": "4:43:46", "remaining_time": "4:16:41"} +{"current_steps": 1975, "total_steps": 3752, "loss": 0.4401, "lr": 2.167306230457564e-05, "epoch": 3.6847014925373136, "percentage": 52.64, "elapsed_time": "4:44:37", "remaining_time": "4:16:05"} +{"current_steps": 1980, "total_steps": 3752, "loss": 0.4749, "lr": 2.1580314044405966e-05, "epoch": 3.6940298507462686, "percentage": 52.77, "elapsed_time": "4:45:09", "remaining_time": "4:15:12"} +{"current_steps": 1985, "total_steps": 3752, "loss": 0.4819, "lr": 2.1487531572312626e-05, "epoch": 3.703358208955224, "percentage": 52.91, "elapsed_time": "4:45:47", "remaining_time": "4:14:24"} +{"current_steps": 1990, "total_steps": 3752, "loss": 0.4421, "lr": 2.139471689692602e-05, "epoch": 3.7126865671641793, "percentage": 53.04, "elapsed_time": "4:46:28", "remaining_time": "4:13:39"} +{"current_steps": 1995, "total_steps": 3752, "loss": 0.4929, "lr": 2.130187202757375e-05, "epoch": 3.7220149253731343, "percentage": 53.17, "elapsed_time": "4:47:18", "remaining_time": "4:13:01"} +{"current_steps": 2000, "total_steps": 3752, "loss": 0.4878, "lr": 2.1208998974237036e-05, "epoch": 3.7313432835820897, "percentage": 53.3, "elapsed_time": "4:48:01", "remaining_time": "4:12:18"} +{"current_steps": 2005, "total_steps": 3752, "loss": 0.4467, "lr": 2.1116099747507274e-05, "epoch": 3.7406716417910446, "percentage": 53.44, "elapsed_time": "4:48:48", "remaining_time": "4:11:39"} +{"current_steps": 2010, "total_steps": 3752, "loss": 0.4796, "lr": 2.1023176358542484e-05, "epoch": 3.75, "percentage": 53.57, "elapsed_time": "4:49:35", "remaining_time": "4:10:58"} +{"current_steps": 2015, "total_steps": 3752, "loss": 0.4162, "lr": 2.093023081902375e-05, "epoch": 3.7593283582089554, "percentage": 53.7, "elapsed_time": "4:50:17", "remaining_time": "4:10:14"} +{"current_steps": 2020, "total_steps": 3752, "loss": 0.4347, "lr": 2.083726514111171e-05, "epoch": 3.7686567164179103, "percentage": 53.84, "elapsed_time": "4:51:02", "remaining_time": "4:09:32"} +{"current_steps": 2025, "total_steps": 3752, "loss": 0.4171, "lr": 2.0744281337402966e-05, "epoch": 3.7779850746268657, "percentage": 53.97, "elapsed_time": "4:51:43", "remaining_time": "4:08:47"} +{"current_steps": 2030, "total_steps": 3752, "loss": 0.4405, "lr": 2.0651281420886522e-05, "epoch": 3.7873134328358207, "percentage": 54.1, "elapsed_time": "4:52:33", "remaining_time": "4:08:10"} +{"current_steps": 2035, "total_steps": 3752, "loss": 0.4841, "lr": 2.0558267404900206e-05, "epoch": 3.796641791044776, "percentage": 54.24, "elapsed_time": "4:53:18", "remaining_time": "4:07:28"} +{"current_steps": 2040, "total_steps": 3752, "loss": 0.4707, "lr": 2.0465241303087078e-05, "epoch": 3.8059701492537314, "percentage": 54.37, "elapsed_time": "4:54:06", "remaining_time": "4:06:49"} +{"current_steps": 2045, "total_steps": 3752, "loss": 0.4334, "lr": 2.0372205129351856e-05, "epoch": 3.8152985074626864, "percentage": 54.5, "elapsed_time": "4:54:51", "remaining_time": "4:06:07"} +{"current_steps": 2050, "total_steps": 3752, "loss": 0.4553, "lr": 2.0279160897817283e-05, "epoch": 3.824626865671642, "percentage": 54.64, "elapsed_time": "4:55:35", "remaining_time": "4:05:24"} +{"current_steps": 2055, "total_steps": 3752, "loss": 0.4481, "lr": 2.0186110622780558e-05, "epoch": 3.833955223880597, "percentage": 54.77, "elapsed_time": "4:56:23", "remaining_time": "4:04:45"} +{"current_steps": 2060, "total_steps": 3752, "loss": 0.4411, "lr": 2.0093056318669718e-05, "epoch": 3.843283582089552, "percentage": 54.9, "elapsed_time": "4:57:18", "remaining_time": "4:04:11"} +{"current_steps": 2065, "total_steps": 3752, "loss": 0.488, "lr": 2e-05, "epoch": 3.8526119402985075, "percentage": 55.04, "elapsed_time": "4:57:59", "remaining_time": "4:03:26"} +{"current_steps": 2070, "total_steps": 3752, "loss": 0.4782, "lr": 1.990694368133029e-05, "epoch": 3.861940298507463, "percentage": 55.17, "elapsed_time": "4:58:35", "remaining_time": "4:02:37"} +{"current_steps": 2075, "total_steps": 3752, "loss": 0.4561, "lr": 1.981388937721945e-05, "epoch": 3.871268656716418, "percentage": 55.3, "elapsed_time": "4:59:18", "remaining_time": "4:01:54"} +{"current_steps": 2080, "total_steps": 3752, "loss": 0.4576, "lr": 1.9720839102182717e-05, "epoch": 3.8805970149253732, "percentage": 55.44, "elapsed_time": "4:59:50", "remaining_time": "4:01:01"} +{"current_steps": 2085, "total_steps": 3752, "loss": 0.4903, "lr": 1.962779487064815e-05, "epoch": 3.8899253731343286, "percentage": 55.57, "elapsed_time": "5:00:43", "remaining_time": "4:00:26"} +{"current_steps": 2090, "total_steps": 3752, "loss": 0.4699, "lr": 1.9534758696912922e-05, "epoch": 3.8992537313432836, "percentage": 55.7, "elapsed_time": "5:01:24", "remaining_time": "3:59:40"} +{"current_steps": 2095, "total_steps": 3752, "loss": 0.4733, "lr": 1.9441732595099804e-05, "epoch": 3.908582089552239, "percentage": 55.84, "elapsed_time": "5:02:12", "remaining_time": "3:59:01"} +{"current_steps": 2100, "total_steps": 3752, "loss": 0.4907, "lr": 1.9348718579113484e-05, "epoch": 3.917910447761194, "percentage": 55.97, "elapsed_time": "5:02:49", "remaining_time": "3:58:13"} +{"current_steps": 2105, "total_steps": 3752, "loss": 0.426, "lr": 1.9255718662597044e-05, "epoch": 3.9272388059701493, "percentage": 56.1, "elapsed_time": "5:03:35", "remaining_time": "3:57:32"} +{"current_steps": 2110, "total_steps": 3752, "loss": 0.4605, "lr": 1.9162734858888293e-05, "epoch": 3.9365671641791042, "percentage": 56.24, "elapsed_time": "5:04:16", "remaining_time": "3:56:47"} +{"current_steps": 2115, "total_steps": 3752, "loss": 0.4309, "lr": 1.9069769180976263e-05, "epoch": 3.9458955223880596, "percentage": 56.37, "elapsed_time": "5:05:06", "remaining_time": "3:56:09"} +{"current_steps": 2120, "total_steps": 3752, "loss": 0.4691, "lr": 1.8976823641457523e-05, "epoch": 3.955223880597015, "percentage": 56.5, "elapsed_time": "5:05:41", "remaining_time": "3:55:19"} +{"current_steps": 2125, "total_steps": 3752, "loss": 0.4808, "lr": 1.888390025249273e-05, "epoch": 3.96455223880597, "percentage": 56.64, "elapsed_time": "5:06:22", "remaining_time": "3:54:34"} +{"current_steps": 2130, "total_steps": 3752, "loss": 0.4525, "lr": 1.879100102576297e-05, "epoch": 3.9738805970149254, "percentage": 56.77, "elapsed_time": "5:07:01", "remaining_time": "3:53:47"} +{"current_steps": 2135, "total_steps": 3752, "loss": 0.4697, "lr": 1.8698127972426255e-05, "epoch": 3.9832089552238807, "percentage": 56.9, "elapsed_time": "5:07:42", "remaining_time": "3:53:03"} +{"current_steps": 2140, "total_steps": 3752, "loss": 0.4348, "lr": 1.8605283103073982e-05, "epoch": 3.9925373134328357, "percentage": 57.04, "elapsed_time": "5:08:26", "remaining_time": "3:52:20"} +{"current_steps": 2145, "total_steps": 3752, "loss": 0.4435, "lr": 1.851246842768738e-05, "epoch": 4.001865671641791, "percentage": 57.17, "elapsed_time": "5:09:11", "remaining_time": "3:51:38"} +{"current_steps": 2150, "total_steps": 3752, "loss": 0.4323, "lr": 1.841968595559404e-05, "epoch": 4.0111940298507465, "percentage": 57.3, "elapsed_time": "5:09:57", "remaining_time": "3:50:57"} +{"current_steps": 2155, "total_steps": 3752, "loss": 0.4378, "lr": 1.8326937695424367e-05, "epoch": 4.020522388059701, "percentage": 57.44, "elapsed_time": "5:10:52", "remaining_time": "3:50:22"} +{"current_steps": 2160, "total_steps": 3752, "loss": 0.4391, "lr": 1.8234225655068133e-05, "epoch": 4.029850746268656, "percentage": 57.57, "elapsed_time": "5:11:31", "remaining_time": "3:49:36"} +{"current_steps": 2165, "total_steps": 3752, "loss": 0.4292, "lr": 1.8141551841630976e-05, "epoch": 4.039179104477612, "percentage": 57.7, "elapsed_time": "5:12:18", "remaining_time": "3:48:55"} +{"current_steps": 2170, "total_steps": 3752, "loss": 0.4263, "lr": 1.804891826139099e-05, "epoch": 4.048507462686567, "percentage": 57.84, "elapsed_time": "5:13:00", "remaining_time": "3:48:11"} +{"current_steps": 2175, "total_steps": 3752, "loss": 0.442, "lr": 1.795632691975526e-05, "epoch": 4.057835820895522, "percentage": 57.97, "elapsed_time": "5:13:39", "remaining_time": "3:47:24"} +{"current_steps": 2180, "total_steps": 3752, "loss": 0.4644, "lr": 1.7863779821216443e-05, "epoch": 4.067164179104478, "percentage": 58.1, "elapsed_time": "5:14:21", "remaining_time": "3:46:40"} +{"current_steps": 2185, "total_steps": 3752, "loss": 0.4248, "lr": 1.77712789693094e-05, "epoch": 4.076492537313433, "percentage": 58.24, "elapsed_time": "5:14:54", "remaining_time": "3:45:50"} +{"current_steps": 2190, "total_steps": 3752, "loss": 0.4503, "lr": 1.76788263665678e-05, "epoch": 4.085820895522388, "percentage": 58.37, "elapsed_time": "5:15:37", "remaining_time": "3:45:06"} +{"current_steps": 2195, "total_steps": 3752, "loss": 0.4511, "lr": 1.758642401448078e-05, "epoch": 4.095149253731344, "percentage": 58.5, "elapsed_time": "5:16:22", "remaining_time": "3:44:24"} +{"current_steps": 2200, "total_steps": 3752, "loss": 0.4513, "lr": 1.74940739134496e-05, "epoch": 4.104477611940299, "percentage": 58.64, "elapsed_time": "5:17:07", "remaining_time": "3:43:43"} +{"current_steps": 2205, "total_steps": 3752, "loss": 0.4478, "lr": 1.7401778062744365e-05, "epoch": 4.1138059701492535, "percentage": 58.77, "elapsed_time": "5:17:38", "remaining_time": "3:42:51"} +{"current_steps": 2210, "total_steps": 3752, "loss": 0.4275, "lr": 1.7309538460460698e-05, "epoch": 4.123134328358209, "percentage": 58.9, "elapsed_time": "5:18:23", "remaining_time": "3:42:09"} +{"current_steps": 2215, "total_steps": 3752, "loss": 0.437, "lr": 1.721735710347654e-05, "epoch": 4.132462686567164, "percentage": 59.04, "elapsed_time": "5:18:59", "remaining_time": "3:41:21"} +{"current_steps": 2220, "total_steps": 3752, "loss": 0.4579, "lr": 1.7125235987408867e-05, "epoch": 4.141791044776119, "percentage": 59.17, "elapsed_time": "5:19:41", "remaining_time": "3:40:36"} +{"current_steps": 2225, "total_steps": 3752, "loss": 0.4611, "lr": 1.7033177106570516e-05, "epoch": 4.151119402985074, "percentage": 59.3, "elapsed_time": "5:20:18", "remaining_time": "3:39:49"} +{"current_steps": 2230, "total_steps": 3752, "loss": 0.437, "lr": 1.694118245392703e-05, "epoch": 4.16044776119403, "percentage": 59.43, "elapsed_time": "5:20:51", "remaining_time": "3:38:59"} +{"current_steps": 2235, "total_steps": 3752, "loss": 0.4305, "lr": 1.684925402105344e-05, "epoch": 4.169776119402985, "percentage": 59.57, "elapsed_time": "5:21:40", "remaining_time": "3:38:19"} +{"current_steps": 2240, "total_steps": 3752, "loss": 0.4531, "lr": 1.6757393798091245e-05, "epoch": 4.17910447761194, "percentage": 59.7, "elapsed_time": "5:22:20", "remaining_time": "3:37:34"} +{"current_steps": 2245, "total_steps": 3752, "loss": 0.4587, "lr": 1.6665603773705247e-05, "epoch": 4.188432835820896, "percentage": 59.83, "elapsed_time": "5:22:59", "remaining_time": "3:36:48"} +{"current_steps": 2250, "total_steps": 3752, "loss": 0.4488, "lr": 1.6573885935040554e-05, "epoch": 4.197761194029851, "percentage": 59.97, "elapsed_time": "5:23:34", "remaining_time": "3:36:00"} +{"current_steps": 2255, "total_steps": 3752, "loss": 0.4547, "lr": 1.648224226767952e-05, "epoch": 4.207089552238806, "percentage": 60.1, "elapsed_time": "5:24:10", "remaining_time": "3:35:12"} +{"current_steps": 2260, "total_steps": 3752, "loss": 0.469, "lr": 1.6390674755598798e-05, "epoch": 4.2164179104477615, "percentage": 60.23, "elapsed_time": "5:24:43", "remaining_time": "3:34:22"} +{"current_steps": 2265, "total_steps": 3752, "loss": 0.4293, "lr": 1.6299185381126344e-05, "epoch": 4.225746268656716, "percentage": 60.37, "elapsed_time": "5:25:25", "remaining_time": "3:33:39"} +{"current_steps": 2270, "total_steps": 3752, "loss": 0.4397, "lr": 1.6207776124898557e-05, "epoch": 4.235074626865671, "percentage": 60.5, "elapsed_time": "5:26:19", "remaining_time": "3:33:02"} +{"current_steps": 2275, "total_steps": 3752, "loss": 0.4497, "lr": 1.611644896581736e-05, "epoch": 4.244402985074627, "percentage": 60.63, "elapsed_time": "5:26:56", "remaining_time": "3:32:15"} +{"current_steps": 2280, "total_steps": 3752, "loss": 0.4176, "lr": 1.602520588100735e-05, "epoch": 4.253731343283582, "percentage": 60.77, "elapsed_time": "5:27:35", "remaining_time": "3:31:29"} +{"current_steps": 2285, "total_steps": 3752, "loss": 0.4293, "lr": 1.5934048845773055e-05, "epoch": 4.263059701492537, "percentage": 60.9, "elapsed_time": "5:28:14", "remaining_time": "3:30:44"} +{"current_steps": 2290, "total_steps": 3752, "loss": 0.4314, "lr": 1.5842979833556084e-05, "epoch": 4.272388059701493, "percentage": 61.03, "elapsed_time": "5:28:55", "remaining_time": "3:29:59"} +{"current_steps": 2295, "total_steps": 3752, "loss": 0.3959, "lr": 1.575200081589251e-05, "epoch": 4.281716417910448, "percentage": 61.17, "elapsed_time": "5:29:44", "remaining_time": "3:29:20"} +{"current_steps": 2300, "total_steps": 3752, "loss": 0.4322, "lr": 1.5661113762370063e-05, "epoch": 4.291044776119403, "percentage": 61.3, "elapsed_time": "5:30:25", "remaining_time": "3:28:35"} +{"current_steps": 2305, "total_steps": 3752, "loss": 0.4234, "lr": 1.5570320640585624e-05, "epoch": 4.300373134328359, "percentage": 61.43, "elapsed_time": "5:31:07", "remaining_time": "3:27:51"} +{"current_steps": 2310, "total_steps": 3752, "loss": 0.4109, "lr": 1.5479623416102512e-05, "epoch": 4.309701492537314, "percentage": 61.57, "elapsed_time": "5:31:49", "remaining_time": "3:27:08"} +{"current_steps": 2315, "total_steps": 3752, "loss": 0.4344, "lr": 1.5389024052408026e-05, "epoch": 4.3190298507462686, "percentage": 61.7, "elapsed_time": "5:32:36", "remaining_time": "3:26:27"} +{"current_steps": 2320, "total_steps": 3752, "loss": 0.444, "lr": 1.529852451087085e-05, "epoch": 4.3283582089552235, "percentage": 61.83, "elapsed_time": "5:33:20", "remaining_time": "3:25:45"} +{"current_steps": 2325, "total_steps": 3752, "loss": 0.4123, "lr": 1.5208126750698676e-05, "epoch": 4.337686567164179, "percentage": 61.97, "elapsed_time": "5:34:07", "remaining_time": "3:25:04"} +{"current_steps": 2330, "total_steps": 3752, "loss": 0.4188, "lr": 1.5117832728895734e-05, "epoch": 4.347014925373134, "percentage": 62.1, "elapsed_time": "5:34:54", "remaining_time": "3:24:23"} +{"current_steps": 2335, "total_steps": 3752, "loss": 0.4563, "lr": 1.5027644400220435e-05, "epoch": 4.356343283582089, "percentage": 62.23, "elapsed_time": "5:35:36", "remaining_time": "3:23:39"} +{"current_steps": 2340, "total_steps": 3752, "loss": 0.4652, "lr": 1.4937563717143077e-05, "epoch": 4.365671641791045, "percentage": 62.37, "elapsed_time": "5:36:13", "remaining_time": "3:22:53"} +{"current_steps": 2345, "total_steps": 3752, "loss": 0.4236, "lr": 1.4847592629803538e-05, "epoch": 4.375, "percentage": 62.5, "elapsed_time": "5:36:57", "remaining_time": "3:22:10"} +{"current_steps": 2350, "total_steps": 3752, "loss": 0.4589, "lr": 1.4757733085969095e-05, "epoch": 4.384328358208955, "percentage": 62.63, "elapsed_time": "5:37:30", "remaining_time": "3:21:21"} +{"current_steps": 2355, "total_steps": 3752, "loss": 0.4871, "lr": 1.4667987030992223e-05, "epoch": 4.393656716417911, "percentage": 62.77, "elapsed_time": "5:38:12", "remaining_time": "3:20:37"} +{"current_steps": 2360, "total_steps": 3752, "loss": 0.4555, "lr": 1.4578356407768524e-05, "epoch": 4.402985074626866, "percentage": 62.9, "elapsed_time": "5:38:53", "remaining_time": "3:19:53"} +{"current_steps": 2365, "total_steps": 3752, "loss": 0.4184, "lr": 1.4488843156694614e-05, "epoch": 4.412313432835821, "percentage": 63.03, "elapsed_time": "5:39:44", "remaining_time": "3:19:15"} +{"current_steps": 2370, "total_steps": 3752, "loss": 0.4224, "lr": 1.4399449215626154e-05, "epoch": 4.4216417910447765, "percentage": 63.17, "elapsed_time": "5:40:34", "remaining_time": "3:18:35"} +{"current_steps": 2375, "total_steps": 3752, "loss": 0.4373, "lr": 1.4310176519835885e-05, "epoch": 4.4309701492537314, "percentage": 63.3, "elapsed_time": "5:41:17", "remaining_time": "3:17:52"} +{"current_steps": 2380, "total_steps": 3752, "loss": 0.4314, "lr": 1.422102700197172e-05, "epoch": 4.440298507462686, "percentage": 63.43, "elapsed_time": "5:41:51", "remaining_time": "3:17:04"} +{"current_steps": 2385, "total_steps": 3752, "loss": 0.4314, "lr": 1.4132002592014939e-05, "epoch": 4.449626865671641, "percentage": 63.57, "elapsed_time": "5:42:25", "remaining_time": "3:16:16"} +{"current_steps": 2390, "total_steps": 3752, "loss": 0.4533, "lr": 1.404310521723836e-05, "epoch": 4.458955223880597, "percentage": 63.7, "elapsed_time": "5:43:18", "remaining_time": "3:15:38"} +{"current_steps": 2395, "total_steps": 3752, "loss": 0.4635, "lr": 1.3954336802164658e-05, "epoch": 4.468283582089552, "percentage": 63.83, "elapsed_time": "5:44:04", "remaining_time": "3:14:57"} +{"current_steps": 2400, "total_steps": 3752, "loss": 0.459, "lr": 1.3865699268524668e-05, "epoch": 4.477611940298507, "percentage": 63.97, "elapsed_time": "5:44:50", "remaining_time": "3:14:15"} +{"current_steps": 2405, "total_steps": 3752, "loss": 0.4584, "lr": 1.3777194535215812e-05, "epoch": 4.486940298507463, "percentage": 64.1, "elapsed_time": "5:46:12", "remaining_time": "3:13:54"} +{"current_steps": 2410, "total_steps": 3752, "loss": 0.4226, "lr": 1.3688824518260528e-05, "epoch": 4.496268656716418, "percentage": 64.23, "elapsed_time": "5:46:44", "remaining_time": "3:13:04"} +{"current_steps": 2415, "total_steps": 3752, "loss": 0.4459, "lr": 1.3600591130764825e-05, "epoch": 4.505597014925373, "percentage": 64.37, "elapsed_time": "5:47:33", "remaining_time": "3:12:24"} +{"current_steps": 2420, "total_steps": 3752, "loss": 0.4677, "lr": 1.3512496282876813e-05, "epoch": 4.514925373134329, "percentage": 64.5, "elapsed_time": "5:48:18", "remaining_time": "3:11:42"} +{"current_steps": 2425, "total_steps": 3752, "loss": 0.4517, "lr": 1.3424541881745425e-05, "epoch": 4.524253731343284, "percentage": 64.63, "elapsed_time": "5:49:04", "remaining_time": "3:11:01"} +{"current_steps": 2430, "total_steps": 3752, "loss": 0.4431, "lr": 1.3336729831479062e-05, "epoch": 4.5335820895522385, "percentage": 64.77, "elapsed_time": "5:49:42", "remaining_time": "3:10:15"} +{"current_steps": 2435, "total_steps": 3752, "loss": 0.4366, "lr": 1.32490620331044e-05, "epoch": 4.542910447761194, "percentage": 64.9, "elapsed_time": "5:50:29", "remaining_time": "3:09:33"} +{"current_steps": 2440, "total_steps": 3752, "loss": 0.4321, "lr": 1.3161540384525244e-05, "epoch": 4.552238805970149, "percentage": 65.03, "elapsed_time": "5:51:12", "remaining_time": "3:08:50"} +{"current_steps": 2445, "total_steps": 3752, "loss": 0.465, "lr": 1.3074166780481412e-05, "epoch": 4.561567164179104, "percentage": 65.17, "elapsed_time": "5:51:50", "remaining_time": "3:08:04"} +{"current_steps": 2450, "total_steps": 3752, "loss": 0.436, "lr": 1.2986943112507753e-05, "epoch": 4.57089552238806, "percentage": 65.3, "elapsed_time": "5:52:30", "remaining_time": "3:07:20"} +{"current_steps": 2455, "total_steps": 3752, "loss": 0.445, "lr": 1.2899871268893161e-05, "epoch": 4.580223880597015, "percentage": 65.43, "elapsed_time": "5:53:12", "remaining_time": "3:06:36"} +{"current_steps": 2460, "total_steps": 3752, "loss": 0.4428, "lr": 1.2812953134639725e-05, "epoch": 4.58955223880597, "percentage": 65.57, "elapsed_time": "5:53:55", "remaining_time": "3:05:52"} +{"current_steps": 2465, "total_steps": 3752, "loss": 0.4047, "lr": 1.27261905914219e-05, "epoch": 4.598880597014926, "percentage": 65.7, "elapsed_time": "5:54:37", "remaining_time": "3:05:08"} +{"current_steps": 2470, "total_steps": 3752, "loss": 0.4443, "lr": 1.2639585517545783e-05, "epoch": 4.608208955223881, "percentage": 65.83, "elapsed_time": "5:55:28", "remaining_time": "3:04:30"} +{"current_steps": 2475, "total_steps": 3752, "loss": 0.4293, "lr": 1.2553139787908457e-05, "epoch": 4.617537313432836, "percentage": 65.96, "elapsed_time": "5:56:06", "remaining_time": "3:03:44"} +{"current_steps": 2480, "total_steps": 3752, "loss": 0.4495, "lr": 1.2466855273957372e-05, "epoch": 4.6268656716417915, "percentage": 66.1, "elapsed_time": "5:56:42", "remaining_time": "3:02:57"} +{"current_steps": 2485, "total_steps": 3752, "loss": 0.4221, "lr": 1.2380733843649866e-05, "epoch": 4.6361940298507465, "percentage": 66.23, "elapsed_time": "5:57:29", "remaining_time": "3:02:16"} +{"current_steps": 2490, "total_steps": 3752, "loss": 0.4318, "lr": 1.2294777361412696e-05, "epoch": 4.645522388059701, "percentage": 66.36, "elapsed_time": "5:58:07", "remaining_time": "3:01:30"} +{"current_steps": 2495, "total_steps": 3752, "loss": 0.4012, "lr": 1.2208987688101701e-05, "epoch": 4.654850746268656, "percentage": 66.5, "elapsed_time": "5:58:50", "remaining_time": "3:00:47"} +{"current_steps": 2500, "total_steps": 3752, "loss": 0.4257, "lr": 1.2123366680961493e-05, "epoch": 4.664179104477612, "percentage": 66.63, "elapsed_time": "5:59:26", "remaining_time": "3:00:00"} +{"current_steps": 2505, "total_steps": 3752, "loss": 0.4366, "lr": 1.2037916193585285e-05, "epoch": 4.673507462686567, "percentage": 66.76, "elapsed_time": "6:00:02", "remaining_time": "2:59:13"} +{"current_steps": 2510, "total_steps": 3752, "loss": 0.4135, "lr": 1.1952638075874696e-05, "epoch": 4.682835820895522, "percentage": 66.9, "elapsed_time": "6:00:37", "remaining_time": "2:58:26"} +{"current_steps": 2515, "total_steps": 3752, "loss": 0.4314, "lr": 1.18675341739998e-05, "epoch": 4.692164179104478, "percentage": 67.03, "elapsed_time": "6:01:23", "remaining_time": "2:57:44"} +{"current_steps": 2520, "total_steps": 3752, "loss": 0.4141, "lr": 1.178260633035906e-05, "epoch": 4.701492537313433, "percentage": 67.16, "elapsed_time": "6:02:15", "remaining_time": "2:57:06"} +{"current_steps": 2525, "total_steps": 3752, "loss": 0.4347, "lr": 1.1697856383539507e-05, "epoch": 4.710820895522388, "percentage": 67.3, "elapsed_time": "6:02:56", "remaining_time": "2:56:22"} +{"current_steps": 2530, "total_steps": 3752, "loss": 0.4327, "lr": 1.1613286168276933e-05, "epoch": 4.720149253731344, "percentage": 67.43, "elapsed_time": "6:03:38", "remaining_time": "2:55:38"} +{"current_steps": 2535, "total_steps": 3752, "loss": 0.4533, "lr": 1.1528897515416114e-05, "epoch": 4.729477611940299, "percentage": 67.56, "elapsed_time": "6:04:20", "remaining_time": "2:54:54"} +{"current_steps": 2540, "total_steps": 3752, "loss": 0.4368, "lr": 1.1444692251871253e-05, "epoch": 4.7388059701492535, "percentage": 67.7, "elapsed_time": "6:05:00", "remaining_time": "2:54:09"} +{"current_steps": 2545, "total_steps": 3752, "loss": 0.4402, "lr": 1.1360672200586365e-05, "epoch": 4.7481343283582085, "percentage": 67.83, "elapsed_time": "6:05:36", "remaining_time": "2:53:23"} +{"current_steps": 2550, "total_steps": 3752, "loss": 0.4319, "lr": 1.1276839180495877e-05, "epoch": 4.757462686567164, "percentage": 67.96, "elapsed_time": "6:06:14", "remaining_time": "2:52:38"} +{"current_steps": 2555, "total_steps": 3752, "loss": 0.4603, "lr": 1.1193195006485153e-05, "epoch": 4.766791044776119, "percentage": 68.1, "elapsed_time": "6:06:54", "remaining_time": "2:51:53"} +{"current_steps": 2560, "total_steps": 3752, "loss": 0.4372, "lr": 1.1109741489351325e-05, "epoch": 4.776119402985074, "percentage": 68.23, "elapsed_time": "6:07:49", "remaining_time": "2:51:16"} +{"current_steps": 2565, "total_steps": 3752, "loss": 0.4426, "lr": 1.1026480435763984e-05, "epoch": 4.78544776119403, "percentage": 68.36, "elapsed_time": "6:08:32", "remaining_time": "2:50:33"} +{"current_steps": 2570, "total_steps": 3752, "loss": 0.4399, "lr": 1.0943413648226142e-05, "epoch": 4.794776119402985, "percentage": 68.5, "elapsed_time": "6:09:13", "remaining_time": "2:49:48"} +{"current_steps": 2575, "total_steps": 3752, "loss": 0.4358, "lr": 1.0860542925035187e-05, "epoch": 4.80410447761194, "percentage": 68.63, "elapsed_time": "6:10:00", "remaining_time": "2:49:07"} +{"current_steps": 2580, "total_steps": 3752, "loss": 0.4645, "lr": 1.0777870060243902e-05, "epoch": 4.813432835820896, "percentage": 68.76, "elapsed_time": "6:10:32", "remaining_time": "2:48:19"} +{"current_steps": 2585, "total_steps": 3752, "loss": 0.4278, "lr": 1.0695396843621722e-05, "epoch": 4.822761194029851, "percentage": 68.9, "elapsed_time": "6:11:17", "remaining_time": "2:47:37"} +{"current_steps": 2590, "total_steps": 3752, "loss": 0.4091, "lr": 1.0613125060615899e-05, "epoch": 4.832089552238806, "percentage": 69.03, "elapsed_time": "6:12:01", "remaining_time": "2:46:54"} +{"current_steps": 2595, "total_steps": 3752, "loss": 0.4332, "lr": 1.0531056492312914e-05, "epoch": 4.8414179104477615, "percentage": 69.16, "elapsed_time": "6:12:48", "remaining_time": "2:46:13"} +{"current_steps": 2600, "total_steps": 3752, "loss": 0.4257, "lr": 1.0449192915399878e-05, "epoch": 4.850746268656716, "percentage": 69.3, "elapsed_time": "6:13:30", "remaining_time": "2:45:29"} +{"current_steps": 2605, "total_steps": 3752, "loss": 0.4538, "lr": 1.0367536102126072e-05, "epoch": 4.860074626865671, "percentage": 69.43, "elapsed_time": "6:14:05", "remaining_time": "2:44:42"} +{"current_steps": 2610, "total_steps": 3752, "loss": 0.3927, "lr": 1.0286087820264603e-05, "epoch": 4.869402985074627, "percentage": 69.56, "elapsed_time": "6:14:48", "remaining_time": "2:43:59"} +{"current_steps": 2615, "total_steps": 3752, "loss": 0.4213, "lr": 1.0204849833074122e-05, "epoch": 4.878731343283582, "percentage": 69.7, "elapsed_time": "6:15:31", "remaining_time": "2:43:16"} +{"current_steps": 2620, "total_steps": 3752, "loss": 0.4514, "lr": 1.0123823899260634e-05, "epoch": 4.888059701492537, "percentage": 69.83, "elapsed_time": "6:16:16", "remaining_time": "2:42:34"} +{"current_steps": 2625, "total_steps": 3752, "loss": 0.4231, "lr": 1.0043011772939466e-05, "epoch": 4.897388059701493, "percentage": 69.96, "elapsed_time": "6:16:58", "remaining_time": "2:41:50"} +{"current_steps": 2630, "total_steps": 3752, "loss": 0.4366, "lr": 9.962415203597245e-06, "epoch": 4.906716417910448, "percentage": 70.1, "elapsed_time": "6:17:40", "remaining_time": "2:41:07"} +{"current_steps": 2635, "total_steps": 3752, "loss": 0.4467, "lr": 9.882035936054042e-06, "epoch": 4.916044776119403, "percentage": 70.23, "elapsed_time": "6:18:14", "remaining_time": "2:40:20"} +{"current_steps": 2640, "total_steps": 3752, "loss": 0.4416, "lr": 9.801875710425625e-06, "epoch": 4.925373134328359, "percentage": 70.36, "elapsed_time": "6:19:00", "remaining_time": "2:39:38"} +{"current_steps": 2645, "total_steps": 3752, "loss": 0.4422, "lr": 9.72193626208574e-06, "epoch": 4.934701492537314, "percentage": 70.5, "elapsed_time": "6:19:50", "remaining_time": "2:38:58"} +{"current_steps": 2650, "total_steps": 3752, "loss": 0.4369, "lr": 9.642219321628592e-06, "epoch": 4.9440298507462686, "percentage": 70.63, "elapsed_time": "6:20:30", "remaining_time": "2:38:13"} +{"current_steps": 2655, "total_steps": 3752, "loss": 0.4494, "lr": 9.56272661483133e-06, "epoch": 4.9533582089552235, "percentage": 70.76, "elapsed_time": "6:21:04", "remaining_time": "2:37:27"} +{"current_steps": 2660, "total_steps": 3752, "loss": 0.4403, "lr": 9.483459862616723e-06, "epoch": 4.962686567164179, "percentage": 70.9, "elapsed_time": "6:21:44", "remaining_time": "2:36:42"} +{"current_steps": 2665, "total_steps": 3752, "loss": 0.4161, "lr": 9.404420781015881e-06, "epoch": 4.972014925373134, "percentage": 71.03, "elapsed_time": "6:22:21", "remaining_time": "2:35:57"} +{"current_steps": 2670, "total_steps": 3752, "loss": 0.4598, "lr": 9.325611081131129e-06, "epoch": 4.981343283582089, "percentage": 71.16, "elapsed_time": "6:23:09", "remaining_time": "2:35:16"} +{"current_steps": 2675, "total_steps": 3752, "loss": 0.445, "lr": 9.247032469098953e-06, "epoch": 4.990671641791045, "percentage": 71.3, "elapsed_time": "6:24:01", "remaining_time": "2:34:36"} +{"current_steps": 2680, "total_steps": 3752, "loss": 0.4368, "lr": 9.168686646053046e-06, "epoch": 5.0, "percentage": 71.43, "elapsed_time": "6:24:47", "remaining_time": "2:33:54"} +{"current_steps": 2685, "total_steps": 3752, "loss": 0.4202, "lr": 9.090575308087502e-06, "epoch": 5.009328358208955, "percentage": 71.56, "elapsed_time": "6:25:28", "remaining_time": "2:33:11"} +{"current_steps": 2690, "total_steps": 3752, "loss": 0.4243, "lr": 9.012700146220082e-06, "epoch": 5.018656716417911, "percentage": 71.7, "elapsed_time": "6:26:11", "remaining_time": "2:32:27"} +{"current_steps": 2695, "total_steps": 3752, "loss": 0.4388, "lr": 8.935062846355642e-06, "epoch": 5.027985074626866, "percentage": 71.83, "elapsed_time": "6:26:54", "remaining_time": "2:31:45"} +{"current_steps": 2700, "total_steps": 3752, "loss": 0.3896, "lr": 8.857665089249572e-06, "epoch": 5.037313432835821, "percentage": 71.96, "elapsed_time": "6:27:38", "remaining_time": "2:31:02"} +{"current_steps": 2705, "total_steps": 3752, "loss": 0.4114, "lr": 8.780508550471483e-06, "epoch": 5.0466417910447765, "percentage": 72.09, "elapsed_time": "6:28:22", "remaining_time": "2:30:19"} +{"current_steps": 2710, "total_steps": 3752, "loss": 0.4495, "lr": 8.703594900368866e-06, "epoch": 5.0559701492537314, "percentage": 72.23, "elapsed_time": "6:29:06", "remaining_time": "2:29:36"} +{"current_steps": 2715, "total_steps": 3752, "loss": 0.4394, "lr": 8.62692580403097e-06, "epoch": 5.065298507462686, "percentage": 72.36, "elapsed_time": "6:30:00", "remaining_time": "2:28:57"} +{"current_steps": 2720, "total_steps": 3752, "loss": 0.3963, "lr": 8.550502921252742e-06, "epoch": 5.074626865671641, "percentage": 72.49, "elapsed_time": "6:30:50", "remaining_time": "2:28:17"} +{"current_steps": 2725, "total_steps": 3752, "loss": 0.4255, "lr": 8.474327906498905e-06, "epoch": 5.083955223880597, "percentage": 72.63, "elapsed_time": "6:31:31", "remaining_time": "2:27:33"} +{"current_steps": 2730, "total_steps": 3752, "loss": 0.3935, "lr": 8.398402408868145e-06, "epoch": 5.093283582089552, "percentage": 72.76, "elapsed_time": "6:32:12", "remaining_time": "2:26:49"} +{"current_steps": 2735, "total_steps": 3752, "loss": 0.4395, "lr": 8.322728072057358e-06, "epoch": 5.102611940298507, "percentage": 72.89, "elapsed_time": "6:32:56", "remaining_time": "2:26:06"} +{"current_steps": 2740, "total_steps": 3752, "loss": 0.4127, "lr": 8.247306534326147e-06, "epoch": 5.111940298507463, "percentage": 73.03, "elapsed_time": "6:33:36", "remaining_time": "2:25:22"} +{"current_steps": 2745, "total_steps": 3752, "loss": 0.4143, "lr": 8.172139428461292e-06, "epoch": 5.121268656716418, "percentage": 73.16, "elapsed_time": "6:34:13", "remaining_time": "2:24:37"} +{"current_steps": 2750, "total_steps": 3752, "loss": 0.4492, "lr": 8.097228381741441e-06, "epoch": 5.130597014925373, "percentage": 73.29, "elapsed_time": "6:34:50", "remaining_time": "2:23:52"} +{"current_steps": 2755, "total_steps": 3752, "loss": 0.4117, "lr": 8.022575015901854e-06, "epoch": 5.139925373134329, "percentage": 73.43, "elapsed_time": "6:35:36", "remaining_time": "2:23:09"} +{"current_steps": 2760, "total_steps": 3752, "loss": 0.4148, "lr": 7.948180947099309e-06, "epoch": 5.149253731343284, "percentage": 73.56, "elapsed_time": "6:36:17", "remaining_time": "2:22:25"} +{"current_steps": 2765, "total_steps": 3752, "loss": 0.4188, "lr": 7.874047785877099e-06, "epoch": 5.1585820895522385, "percentage": 73.69, "elapsed_time": "6:36:54", "remaining_time": "2:21:40"} +{"current_steps": 2770, "total_steps": 3752, "loss": 0.4239, "lr": 7.800177137130198e-06, "epoch": 5.167910447761194, "percentage": 73.83, "elapsed_time": "6:37:37", "remaining_time": "2:20:57"} +{"current_steps": 2775, "total_steps": 3752, "loss": 0.4333, "lr": 7.726570600070498e-06, "epoch": 5.177238805970149, "percentage": 73.96, "elapsed_time": "6:38:21", "remaining_time": "2:20:14"} +{"current_steps": 2780, "total_steps": 3752, "loss": 0.4384, "lr": 7.653229768192163e-06, "epoch": 5.186567164179104, "percentage": 74.09, "elapsed_time": "6:38:58", "remaining_time": "2:19:29"} +{"current_steps": 2785, "total_steps": 3752, "loss": 0.4201, "lr": 7.580156229237172e-06, "epoch": 5.19589552238806, "percentage": 74.23, "elapsed_time": "6:39:43", "remaining_time": "2:18:47"} +{"current_steps": 2790, "total_steps": 3752, "loss": 0.424, "lr": 7.507351565160918e-06, "epoch": 5.205223880597015, "percentage": 74.36, "elapsed_time": "6:40:33", "remaining_time": "2:18:06"} +{"current_steps": 2795, "total_steps": 3752, "loss": 0.4112, "lr": 7.434817352097994e-06, "epoch": 5.21455223880597, "percentage": 74.49, "elapsed_time": "6:41:10", "remaining_time": "2:17:21"} +{"current_steps": 2800, "total_steps": 3752, "loss": 0.4317, "lr": 7.36255516032802e-06, "epoch": 5.223880597014926, "percentage": 74.63, "elapsed_time": "6:41:52", "remaining_time": "2:16:38"} +{"current_steps": 2805, "total_steps": 3752, "loss": 0.4193, "lr": 7.290566554241705e-06, "epoch": 5.233208955223881, "percentage": 74.76, "elapsed_time": "6:42:37", "remaining_time": "2:15:55"} +{"current_steps": 2810, "total_steps": 3752, "loss": 0.4121, "lr": 7.218853092306937e-06, "epoch": 5.242537313432836, "percentage": 74.89, "elapsed_time": "6:43:28", "remaining_time": "2:15:15"} +{"current_steps": 2815, "total_steps": 3752, "loss": 0.449, "lr": 7.147416327035057e-06, "epoch": 5.251865671641791, "percentage": 75.03, "elapsed_time": "6:44:06", "remaining_time": "2:14:30"} +{"current_steps": 2820, "total_steps": 3752, "loss": 0.4266, "lr": 7.0762578049472574e-06, "epoch": 5.2611940298507465, "percentage": 75.16, "elapsed_time": "6:44:48", "remaining_time": "2:13:47"} +{"current_steps": 2825, "total_steps": 3752, "loss": 0.4259, "lr": 7.005379066541101e-06, "epoch": 5.270522388059701, "percentage": 75.29, "elapsed_time": "6:45:30", "remaining_time": "2:13:03"} +{"current_steps": 2830, "total_steps": 3752, "loss": 0.4341, "lr": 6.9347816462571675e-06, "epoch": 5.279850746268656, "percentage": 75.43, "elapsed_time": "6:46:22", "remaining_time": "2:12:23"} +{"current_steps": 2835, "total_steps": 3752, "loss": 0.402, "lr": 6.864467072445817e-06, "epoch": 5.289179104477612, "percentage": 75.56, "elapsed_time": "6:47:07", "remaining_time": "2:11:41"} +{"current_steps": 2840, "total_steps": 3752, "loss": 0.4126, "lr": 6.794436867334133e-06, "epoch": 5.298507462686567, "percentage": 75.69, "elapsed_time": "6:47:45", "remaining_time": "2:10:56"} +{"current_steps": 2845, "total_steps": 3752, "loss": 0.4502, "lr": 6.7246925469929395e-06, "epoch": 5.307835820895522, "percentage": 75.83, "elapsed_time": "6:48:33", "remaining_time": "2:10:15"} +{"current_steps": 2850, "total_steps": 3752, "loss": 0.4033, "lr": 6.655235621304021e-06, "epoch": 5.317164179104478, "percentage": 75.96, "elapsed_time": "6:49:14", "remaining_time": "2:09:31"} +{"current_steps": 2855, "total_steps": 3752, "loss": 0.4271, "lr": 6.5860675939273766e-06, "epoch": 5.326492537313433, "percentage": 76.09, "elapsed_time": "6:49:52", "remaining_time": "2:08:46"} +{"current_steps": 2860, "total_steps": 3752, "loss": 0.4097, "lr": 6.517189962268733e-06, "epoch": 5.335820895522388, "percentage": 76.23, "elapsed_time": "6:50:35", "remaining_time": "2:08:03"} +{"current_steps": 2865, "total_steps": 3752, "loss": 0.4641, "lr": 6.448604217447067e-06, "epoch": 5.345149253731344, "percentage": 76.36, "elapsed_time": "6:51:14", "remaining_time": "2:07:19"} +{"current_steps": 2870, "total_steps": 3752, "loss": 0.4144, "lr": 6.3803118442623565e-06, "epoch": 5.354477611940299, "percentage": 76.49, "elapsed_time": "6:52:03", "remaining_time": "2:06:37"} +{"current_steps": 2875, "total_steps": 3752, "loss": 0.4151, "lr": 6.312314321163449e-06, "epoch": 5.3638059701492535, "percentage": 76.63, "elapsed_time": "6:52:50", "remaining_time": "2:05:56"} +{"current_steps": 2880, "total_steps": 3752, "loss": 0.416, "lr": 6.244613120216014e-06, "epoch": 5.373134328358209, "percentage": 76.76, "elapsed_time": "6:53:33", "remaining_time": "2:05:13"} +{"current_steps": 2885, "total_steps": 3752, "loss": 0.3955, "lr": 6.177209707070722e-06, "epoch": 5.382462686567164, "percentage": 76.89, "elapsed_time": "6:54:06", "remaining_time": "2:04:26"} +{"current_steps": 2890, "total_steps": 3752, "loss": 0.449, "lr": 6.110105540931472e-06, "epoch": 5.391791044776119, "percentage": 77.03, "elapsed_time": "6:54:49", "remaining_time": "2:03:43"} +{"current_steps": 2895, "total_steps": 3752, "loss": 0.4107, "lr": 6.043302074523827e-06, "epoch": 5.401119402985074, "percentage": 77.16, "elapsed_time": "6:55:29", "remaining_time": "2:02:59"} +{"current_steps": 2900, "total_steps": 3752, "loss": 0.4343, "lr": 5.976800754063552e-06, "epoch": 5.41044776119403, "percentage": 77.29, "elapsed_time": "6:56:03", "remaining_time": "2:02:14"} +{"current_steps": 2905, "total_steps": 3752, "loss": 0.4613, "lr": 5.9106030192253275e-06, "epoch": 5.419776119402985, "percentage": 77.43, "elapsed_time": "6:56:42", "remaining_time": "2:01:29"} +{"current_steps": 2910, "total_steps": 3752, "loss": 0.437, "lr": 5.844710303111541e-06, "epoch": 5.42910447761194, "percentage": 77.56, "elapsed_time": "6:57:30", "remaining_time": "2:00:48"} +{"current_steps": 2915, "total_steps": 3752, "loss": 0.4181, "lr": 5.7791240322213215e-06, "epoch": 5.438432835820896, "percentage": 77.69, "elapsed_time": "6:58:04", "remaining_time": "2:00:02"} +{"current_steps": 2920, "total_steps": 3752, "loss": 0.3974, "lr": 5.7138456264195765e-06, "epoch": 5.447761194029851, "percentage": 77.83, "elapsed_time": "6:58:47", "remaining_time": "1:59:19"} +{"current_steps": 2925, "total_steps": 3752, "loss": 0.4807, "lr": 5.648876498906335e-06, "epoch": 5.457089552238806, "percentage": 77.96, "elapsed_time": "6:59:24", "remaining_time": "1:58:34"} +{"current_steps": 2930, "total_steps": 3752, "loss": 0.4189, "lr": 5.584218056186113e-06, "epoch": 5.4664179104477615, "percentage": 78.09, "elapsed_time": "7:00:11", "remaining_time": "1:57:53"} +{"current_steps": 2935, "total_steps": 3752, "loss": 0.4279, "lr": 5.519871698037453e-06, "epoch": 5.475746268656716, "percentage": 78.22, "elapsed_time": "7:00:53", "remaining_time": "1:57:09"} +{"current_steps": 2940, "total_steps": 3752, "loss": 0.4577, "lr": 5.455838817482666e-06, "epoch": 5.485074626865671, "percentage": 78.36, "elapsed_time": "7:01:36", "remaining_time": "1:56:26"} +{"current_steps": 2945, "total_steps": 3752, "loss": 0.4297, "lr": 5.392120800757603e-06, "epoch": 5.494402985074627, "percentage": 78.49, "elapsed_time": "7:02:21", "remaining_time": "1:55:44"} +{"current_steps": 2950, "total_steps": 3752, "loss": 0.4251, "lr": 5.328719027281728e-06, "epoch": 5.503731343283582, "percentage": 78.62, "elapsed_time": "7:03:06", "remaining_time": "1:55:01"} +{"current_steps": 2955, "total_steps": 3752, "loss": 0.4377, "lr": 5.265634869628182e-06, "epoch": 5.513059701492537, "percentage": 78.76, "elapsed_time": "7:03:46", "remaining_time": "1:54:17"} +{"current_steps": 2960, "total_steps": 3752, "loss": 0.4153, "lr": 5.20286969349413e-06, "epoch": 5.522388059701493, "percentage": 78.89, "elapsed_time": "7:04:37", "remaining_time": "1:53:36"} +{"current_steps": 2965, "total_steps": 3752, "loss": 0.4108, "lr": 5.140424857671145e-06, "epoch": 5.531716417910448, "percentage": 79.02, "elapsed_time": "7:05:21", "remaining_time": "1:52:54"} +{"current_steps": 2970, "total_steps": 3752, "loss": 0.4204, "lr": 5.07830171401581e-06, "epoch": 5.541044776119403, "percentage": 79.16, "elapsed_time": "7:06:05", "remaining_time": "1:52:11"} +{"current_steps": 2975, "total_steps": 3752, "loss": 0.4157, "lr": 5.016501607420476e-06, "epoch": 5.550373134328359, "percentage": 79.29, "elapsed_time": "7:06:49", "remaining_time": "1:51:28"} +{"current_steps": 2980, "total_steps": 3752, "loss": 0.4395, "lr": 4.955025875784098e-06, "epoch": 5.559701492537314, "percentage": 79.42, "elapsed_time": "7:07:41", "remaining_time": "1:50:47"} +{"current_steps": 2985, "total_steps": 3752, "loss": 0.4223, "lr": 4.893875849983323e-06, "epoch": 5.5690298507462686, "percentage": 79.56, "elapsed_time": "7:08:13", "remaining_time": "1:50:01"} +{"current_steps": 2990, "total_steps": 3752, "loss": 0.4584, "lr": 4.833052853843636e-06, "epoch": 5.5783582089552235, "percentage": 79.69, "elapsed_time": "7:08:58", "remaining_time": "1:49:19"} +{"current_steps": 2995, "total_steps": 3752, "loss": 0.4107, "lr": 4.772558204110724e-06, "epoch": 5.587686567164179, "percentage": 79.82, "elapsed_time": "7:09:40", "remaining_time": "1:48:36"} +{"current_steps": 3000, "total_steps": 3752, "loss": 0.429, "lr": 4.712393210421957e-06, "epoch": 5.597014925373134, "percentage": 79.96, "elapsed_time": "7:10:23", "remaining_time": "1:47:53"} +{"current_steps": 3005, "total_steps": 3752, "loss": 0.4134, "lr": 4.652559175278062e-06, "epoch": 5.606343283582089, "percentage": 80.09, "elapsed_time": "7:12:06", "remaining_time": "1:47:24"} +{"current_steps": 3010, "total_steps": 3752, "loss": 0.3873, "lr": 4.593057394014882e-06, "epoch": 5.615671641791045, "percentage": 80.22, "elapsed_time": "7:12:53", "remaining_time": "1:46:42"} +{"current_steps": 3015, "total_steps": 3752, "loss": 0.403, "lr": 4.5338891547753775e-06, "epoch": 5.625, "percentage": 80.36, "elapsed_time": "7:13:42", "remaining_time": "1:46:01"} +{"current_steps": 3020, "total_steps": 3752, "loss": 0.4058, "lr": 4.475055738481711e-06, "epoch": 5.634328358208955, "percentage": 80.49, "elapsed_time": "7:14:26", "remaining_time": "1:45:18"} +{"current_steps": 3025, "total_steps": 3752, "loss": 0.4497, "lr": 4.416558418807517e-06, "epoch": 5.643656716417911, "percentage": 80.62, "elapsed_time": "7:15:01", "remaining_time": "1:44:32"} +{"current_steps": 3030, "total_steps": 3752, "loss": 0.4126, "lr": 4.3583984621503596e-06, "epoch": 5.652985074626866, "percentage": 80.76, "elapsed_time": "7:15:45", "remaining_time": "1:43:50"} +{"current_steps": 3035, "total_steps": 3752, "loss": 0.4234, "lr": 4.300577127604269e-06, "epoch": 5.662313432835821, "percentage": 80.89, "elapsed_time": "7:16:25", "remaining_time": "1:43:06"} +{"current_steps": 3040, "total_steps": 3752, "loss": 0.4259, "lr": 4.243095666932531e-06, "epoch": 5.6716417910447765, "percentage": 81.02, "elapsed_time": "7:17:09", "remaining_time": "1:42:23"} +{"current_steps": 3045, "total_steps": 3752, "loss": 0.4191, "lr": 4.18595532454055e-06, "epoch": 5.6809701492537314, "percentage": 81.16, "elapsed_time": "7:17:46", "remaining_time": "1:41:38"} +{"current_steps": 3050, "total_steps": 3752, "loss": 0.4317, "lr": 4.129157337448926e-06, "epoch": 5.690298507462686, "percentage": 81.29, "elapsed_time": "7:18:29", "remaining_time": "1:40:55"} +{"current_steps": 3055, "total_steps": 3752, "loss": 0.4047, "lr": 4.072702935266677e-06, "epoch": 5.699626865671641, "percentage": 81.42, "elapsed_time": "7:19:07", "remaining_time": "1:40:11"} +{"current_steps": 3060, "total_steps": 3752, "loss": 0.4091, "lr": 4.016593340164618e-06, "epoch": 5.708955223880597, "percentage": 81.56, "elapsed_time": "7:19:53", "remaining_time": "1:39:28"} +{"current_steps": 3065, "total_steps": 3752, "loss": 0.4141, "lr": 3.960829766848893e-06, "epoch": 5.718283582089552, "percentage": 81.69, "elapsed_time": "7:20:42", "remaining_time": "1:38:47"} +{"current_steps": 3070, "total_steps": 3752, "loss": 0.4296, "lr": 3.905413422534696e-06, "epoch": 5.727611940298507, "percentage": 81.82, "elapsed_time": "7:21:24", "remaining_time": "1:38:03"} +{"current_steps": 3075, "total_steps": 3752, "loss": 0.3867, "lr": 3.85034550692011e-06, "epoch": 5.736940298507463, "percentage": 81.96, "elapsed_time": "7:22:11", "remaining_time": "1:37:21"} +{"current_steps": 3080, "total_steps": 3752, "loss": 0.3884, "lr": 3.7956272121601555e-06, "epoch": 5.746268656716418, "percentage": 82.09, "elapsed_time": "7:22:49", "remaining_time": "1:36:37"} +{"current_steps": 3085, "total_steps": 3752, "loss": 0.4404, "lr": 3.7412597228409884e-06, "epoch": 5.755597014925373, "percentage": 82.22, "elapsed_time": "7:23:22", "remaining_time": "1:35:51"} +{"current_steps": 3090, "total_steps": 3752, "loss": 0.4433, "lr": 3.687244215954222e-06, "epoch": 5.764925373134329, "percentage": 82.36, "elapsed_time": "7:24:08", "remaining_time": "1:35:09"} +{"current_steps": 3095, "total_steps": 3752, "loss": 0.4044, "lr": 3.633581860871491e-06, "epoch": 5.774253731343284, "percentage": 82.49, "elapsed_time": "7:24:56", "remaining_time": "1:34:27"} +{"current_steps": 3100, "total_steps": 3752, "loss": 0.4356, "lr": 3.5802738193190954e-06, "epoch": 5.7835820895522385, "percentage": 82.62, "elapsed_time": "7:25:39", "remaining_time": "1:33:43"} +{"current_steps": 3105, "total_steps": 3752, "loss": 0.3999, "lr": 3.5273212453528705e-06, "epoch": 5.792910447761194, "percentage": 82.76, "elapsed_time": "7:26:21", "remaining_time": "1:33:00"} +{"current_steps": 3110, "total_steps": 3752, "loss": 0.4334, "lr": 3.4747252853332004e-06, "epoch": 5.802238805970149, "percentage": 82.89, "elapsed_time": "7:26:57", "remaining_time": "1:32:15"} +{"current_steps": 3115, "total_steps": 3752, "loss": 0.3906, "lr": 3.4224870779002117e-06, "epoch": 5.811567164179104, "percentage": 83.02, "elapsed_time": "7:27:45", "remaining_time": "1:31:33"} +{"current_steps": 3120, "total_steps": 3752, "loss": 0.4465, "lr": 3.3706077539490933e-06, "epoch": 5.82089552238806, "percentage": 83.16, "elapsed_time": "7:28:24", "remaining_time": "1:30:49"} +{"current_steps": 3125, "total_steps": 3752, "loss": 0.4292, "lr": 3.3190884366056532e-06, "epoch": 5.830223880597015, "percentage": 83.29, "elapsed_time": "7:29:12", "remaining_time": "1:30:07"} +{"current_steps": 3130, "total_steps": 3752, "loss": 0.3936, "lr": 3.2679302412019665e-06, "epoch": 5.83955223880597, "percentage": 83.42, "elapsed_time": "7:29:56", "remaining_time": "1:29:24"} +{"current_steps": 3135, "total_steps": 3752, "loss": 0.4604, "lr": 3.2171342752522494e-06, "epoch": 5.848880597014926, "percentage": 83.56, "elapsed_time": "7:30:31", "remaining_time": "1:28:40"} +{"current_steps": 3140, "total_steps": 3752, "loss": 0.4328, "lr": 3.166701638428895e-06, "epoch": 5.858208955223881, "percentage": 83.69, "elapsed_time": "7:31:12", "remaining_time": "1:27:56"} +{"current_steps": 3145, "total_steps": 3752, "loss": 0.4053, "lr": 3.1166334225386306e-06, "epoch": 5.867537313432836, "percentage": 83.82, "elapsed_time": "7:31:50", "remaining_time": "1:27:12"} +{"current_steps": 3150, "total_steps": 3752, "loss": 0.3972, "lr": 3.066930711498921e-06, "epoch": 5.8768656716417915, "percentage": 83.96, "elapsed_time": "7:32:27", "remaining_time": "1:26:28"} +{"current_steps": 3155, "total_steps": 3752, "loss": 0.4064, "lr": 3.0175945813144668e-06, "epoch": 5.8861940298507465, "percentage": 84.09, "elapsed_time": "7:33:14", "remaining_time": "1:25:45"} +{"current_steps": 3160, "total_steps": 3752, "loss": 0.4655, "lr": 2.9686261000539484e-06, "epoch": 5.895522388059701, "percentage": 84.22, "elapsed_time": "7:33:54", "remaining_time": "1:25:02"} +{"current_steps": 3165, "total_steps": 3752, "loss": 0.4079, "lr": 2.920026327826866e-06, "epoch": 5.904850746268656, "percentage": 84.36, "elapsed_time": "7:34:42", "remaining_time": "1:24:20"} +{"current_steps": 3170, "total_steps": 3752, "loss": 0.4253, "lr": 2.87179631676062e-06, "epoch": 5.914179104477612, "percentage": 84.49, "elapsed_time": "7:35:15", "remaining_time": "1:23:34"} +{"current_steps": 3175, "total_steps": 3752, "loss": 0.4172, "lr": 2.8239371109777127e-06, "epoch": 5.923507462686567, "percentage": 84.62, "elapsed_time": "7:35:57", "remaining_time": "1:22:51"} +{"current_steps": 3180, "total_steps": 3752, "loss": 0.4288, "lr": 2.7764497465731487e-06, "epoch": 5.932835820895522, "percentage": 84.75, "elapsed_time": "7:36:44", "remaining_time": "1:22:09"} +{"current_steps": 3185, "total_steps": 3752, "loss": 0.4133, "lr": 2.72933525159202e-06, "epoch": 5.942164179104478, "percentage": 84.89, "elapsed_time": "7:37:27", "remaining_time": "1:21:26"} +{"current_steps": 3190, "total_steps": 3752, "loss": 0.3976, "lr": 2.6825946460072237e-06, "epoch": 5.951492537313433, "percentage": 85.02, "elapsed_time": "7:38:14", "remaining_time": "1:20:43"} +{"current_steps": 3195, "total_steps": 3752, "loss": 0.4335, "lr": 2.636228941697414e-06, "epoch": 5.960820895522388, "percentage": 85.15, "elapsed_time": "7:38:58", "remaining_time": "1:20:00"} +{"current_steps": 3200, "total_steps": 3752, "loss": 0.3967, "lr": 2.5902391424250573e-06, "epoch": 5.970149253731344, "percentage": 85.29, "elapsed_time": "7:39:35", "remaining_time": "1:19:16"} +{"current_steps": 3205, "total_steps": 3752, "loss": 0.414, "lr": 2.544626243814732e-06, "epoch": 5.979477611940299, "percentage": 85.42, "elapsed_time": "7:40:18", "remaining_time": "1:18:33"} +{"current_steps": 3210, "total_steps": 3752, "loss": 0.4365, "lr": 2.4993912333315605e-06, "epoch": 5.9888059701492535, "percentage": 85.55, "elapsed_time": "7:41:00", "remaining_time": "1:17:50"} +{"current_steps": 3215, "total_steps": 3752, "loss": 0.4341, "lr": 2.4545350902598464e-06, "epoch": 5.9981343283582085, "percentage": 85.69, "elapsed_time": "7:41:37", "remaining_time": "1:17:06"} +{"current_steps": 3220, "total_steps": 3752, "loss": 0.3946, "lr": 2.4100587856818458e-06, "epoch": 6.007462686567164, "percentage": 85.82, "elapsed_time": "7:42:24", "remaining_time": "1:16:23"} +{"current_steps": 3225, "total_steps": 3752, "loss": 0.404, "lr": 2.3659632824567823e-06, "epoch": 6.016791044776119, "percentage": 85.95, "elapsed_time": "7:43:08", "remaining_time": "1:15:40"} +{"current_steps": 3230, "total_steps": 3752, "loss": 0.4044, "lr": 2.322249535199965e-06, "epoch": 6.026119402985074, "percentage": 86.09, "elapsed_time": "7:43:42", "remaining_time": "1:14:56"} +{"current_steps": 3235, "total_steps": 3752, "loss": 0.4357, "lr": 2.278918490262143e-06, "epoch": 6.03544776119403, "percentage": 86.22, "elapsed_time": "7:44:23", "remaining_time": "1:14:13"} +{"current_steps": 3240, "total_steps": 3752, "loss": 0.4029, "lr": 2.235971085709028e-06, "epoch": 6.044776119402985, "percentage": 86.35, "elapsed_time": "7:45:08", "remaining_time": "1:13:30"} +{"current_steps": 3245, "total_steps": 3752, "loss": 0.4507, "lr": 2.193408251300948e-06, "epoch": 6.05410447761194, "percentage": 86.49, "elapsed_time": "7:45:54", "remaining_time": "1:12:47"} +{"current_steps": 3250, "total_steps": 3752, "loss": 0.4056, "lr": 2.1512309084727724e-06, "epoch": 6.063432835820896, "percentage": 86.62, "elapsed_time": "7:46:37", "remaining_time": "1:12:04"} +{"current_steps": 3255, "total_steps": 3752, "loss": 0.406, "lr": 2.1094399703139113e-06, "epoch": 6.072761194029851, "percentage": 86.75, "elapsed_time": "7:47:28", "remaining_time": "1:11:22"} +{"current_steps": 3260, "total_steps": 3752, "loss": 0.4132, "lr": 2.068036341548585e-06, "epoch": 6.082089552238806, "percentage": 86.89, "elapsed_time": "7:48:09", "remaining_time": "1:10:39"} +{"current_steps": 3265, "total_steps": 3752, "loss": 0.426, "lr": 2.027020918516216e-06, "epoch": 6.0914179104477615, "percentage": 87.02, "elapsed_time": "7:48:53", "remaining_time": "1:09:56"} +{"current_steps": 3270, "total_steps": 3752, "loss": 0.4323, "lr": 1.9863945891520474e-06, "epoch": 6.100746268656716, "percentage": 87.15, "elapsed_time": "7:49:26", "remaining_time": "1:09:11"} +{"current_steps": 3275, "total_steps": 3752, "loss": 0.4571, "lr": 1.9461582329678895e-06, "epoch": 6.110074626865671, "percentage": 87.29, "elapsed_time": "7:50:05", "remaining_time": "1:08:28"} +{"current_steps": 3280, "total_steps": 3752, "loss": 0.4423, "lr": 1.9063127210331145e-06, "epoch": 6.119402985074627, "percentage": 87.42, "elapsed_time": "7:50:46", "remaining_time": "1:07:44"} +{"current_steps": 3285, "total_steps": 3752, "loss": 0.4205, "lr": 1.8668589159557626e-06, "epoch": 6.128731343283582, "percentage": 87.55, "elapsed_time": "7:51:34", "remaining_time": "1:07:02"} +{"current_steps": 3290, "total_steps": 3752, "loss": 0.4174, "lr": 1.8277976718638934e-06, "epoch": 6.138059701492537, "percentage": 87.69, "elapsed_time": "7:52:11", "remaining_time": "1:06:18"} +{"current_steps": 3295, "total_steps": 3752, "loss": 0.418, "lr": 1.789129834387091e-06, "epoch": 6.147388059701493, "percentage": 87.82, "elapsed_time": "7:53:02", "remaining_time": "1:05:36"} +{"current_steps": 3300, "total_steps": 3752, "loss": 0.407, "lr": 1.7508562406381414e-06, "epoch": 6.156716417910448, "percentage": 87.95, "elapsed_time": "7:53:40", "remaining_time": "1:04:52"} +{"current_steps": 3305, "total_steps": 3752, "loss": 0.3783, "lr": 1.7129777191949437e-06, "epoch": 6.166044776119403, "percentage": 88.09, "elapsed_time": "7:54:30", "remaining_time": "1:04:10"} +{"current_steps": 3310, "total_steps": 3752, "loss": 0.419, "lr": 1.6754950900825174e-06, "epoch": 6.175373134328359, "percentage": 88.22, "elapsed_time": "7:55:04", "remaining_time": "1:03:26"} +{"current_steps": 3315, "total_steps": 3752, "loss": 0.437, "lr": 1.6384091647553125e-06, "epoch": 6.184701492537314, "percentage": 88.35, "elapsed_time": "7:55:44", "remaining_time": "1:02:42"} +{"current_steps": 3320, "total_steps": 3752, "loss": 0.3872, "lr": 1.601720746079598e-06, "epoch": 6.1940298507462686, "percentage": 88.49, "elapsed_time": "7:56:23", "remaining_time": "1:01:59"} +{"current_steps": 3325, "total_steps": 3752, "loss": 0.3896, "lr": 1.5654306283161002e-06, "epoch": 6.2033582089552235, "percentage": 88.62, "elapsed_time": "7:56:59", "remaining_time": "1:01:15"} +{"current_steps": 3330, "total_steps": 3752, "loss": 0.4202, "lr": 1.5295395971028094e-06, "epoch": 6.212686567164179, "percentage": 88.75, "elapsed_time": "7:57:40", "remaining_time": "1:00:32"} +{"current_steps": 3335, "total_steps": 3752, "loss": 0.4335, "lr": 1.4940484294379442e-06, "epoch": 6.222014925373134, "percentage": 88.89, "elapsed_time": "7:58:23", "remaining_time": "0:59:49"} +{"current_steps": 3340, "total_steps": 3752, "loss": 0.4131, "lr": 1.4589578936631776e-06, "epoch": 6.231343283582089, "percentage": 89.02, "elapsed_time": "7:59:10", "remaining_time": "0:59:06"} +{"current_steps": 3345, "total_steps": 3752, "loss": 0.3663, "lr": 1.424268749446962e-06, "epoch": 6.240671641791045, "percentage": 89.15, "elapsed_time": "7:59:58", "remaining_time": "0:58:24"} +{"current_steps": 3350, "total_steps": 3752, "loss": 0.4162, "lr": 1.3899817477681056e-06, "epoch": 6.25, "percentage": 89.29, "elapsed_time": "8:00:34", "remaining_time": "0:57:40"} +{"current_steps": 3355, "total_steps": 3752, "loss": 0.3911, "lr": 1.3560976308995065e-06, "epoch": 6.259328358208955, "percentage": 89.42, "elapsed_time": "8:01:18", "remaining_time": "0:56:57"} +{"current_steps": 3360, "total_steps": 3752, "loss": 0.4385, "lr": 1.3226171323920811e-06, "epoch": 6.268656716417911, "percentage": 89.55, "elapsed_time": "8:01:57", "remaining_time": "0:56:13"} +{"current_steps": 3365, "total_steps": 3752, "loss": 0.3937, "lr": 1.2895409770588874e-06, "epoch": 6.277985074626866, "percentage": 89.69, "elapsed_time": "8:02:36", "remaining_time": "0:55:30"} +{"current_steps": 3370, "total_steps": 3752, "loss": 0.4036, "lr": 1.2568698809594437e-06, "epoch": 6.287313432835821, "percentage": 89.82, "elapsed_time": "8:03:13", "remaining_time": "0:54:46"} +{"current_steps": 3375, "total_steps": 3752, "loss": 0.3823, "lr": 1.2246045513842008e-06, "epoch": 6.2966417910447765, "percentage": 89.95, "elapsed_time": "8:03:56", "remaining_time": "0:54:03"} +{"current_steps": 3380, "total_steps": 3752, "loss": 0.4252, "lr": 1.1927456868392605e-06, "epoch": 6.3059701492537314, "percentage": 90.09, "elapsed_time": "8:04:40", "remaining_time": "0:53:20"} +{"current_steps": 3385, "total_steps": 3752, "loss": 0.4473, "lr": 1.1612939770312325e-06, "epoch": 6.315298507462686, "percentage": 90.22, "elapsed_time": "8:05:27", "remaining_time": "0:52:37"} +{"current_steps": 3390, "total_steps": 3752, "loss": 0.4836, "lr": 1.1302501028523039e-06, "epoch": 6.324626865671641, "percentage": 90.35, "elapsed_time": "8:06:09", "remaining_time": "0:51:54"} +{"current_steps": 3395, "total_steps": 3752, "loss": 0.3809, "lr": 1.0996147363655175e-06, "epoch": 6.333955223880597, "percentage": 90.49, "elapsed_time": "8:06:51", "remaining_time": "0:51:11"} +{"current_steps": 3400, "total_steps": 3752, "loss": 0.4204, "lr": 1.069388540790195e-06, "epoch": 6.343283582089552, "percentage": 90.62, "elapsed_time": "8:07:34", "remaining_time": "0:50:28"} +{"current_steps": 3405, "total_steps": 3752, "loss": 0.3976, "lr": 1.039572170487606e-06, "epoch": 6.352611940298507, "percentage": 90.75, "elapsed_time": "8:08:10", "remaining_time": "0:49:45"} +{"current_steps": 3410, "total_steps": 3752, "loss": 0.4114, "lr": 1.0101662709467841e-06, "epoch": 6.361940298507463, "percentage": 90.88, "elapsed_time": "8:08:56", "remaining_time": "0:49:02"} +{"current_steps": 3415, "total_steps": 3752, "loss": 0.4134, "lr": 9.811714787705528e-07, "epoch": 6.371268656716418, "percentage": 91.02, "elapsed_time": "8:09:36", "remaining_time": "0:48:18"} +{"current_steps": 3420, "total_steps": 3752, "loss": 0.4551, "lr": 9.525884216617575e-07, "epoch": 6.380597014925373, "percentage": 91.15, "elapsed_time": "8:10:12", "remaining_time": "0:47:35"} +{"current_steps": 3425, "total_steps": 3752, "loss": 0.4088, "lr": 9.244177184096603e-07, "epoch": 6.389925373134329, "percentage": 91.28, "elapsed_time": "8:10:53", "remaining_time": "0:46:52"} +{"current_steps": 3430, "total_steps": 3752, "loss": 0.4445, "lr": 8.96659978876564e-07, "epoch": 6.399253731343284, "percentage": 91.42, "elapsed_time": "8:11:41", "remaining_time": "0:46:09"} +{"current_steps": 3435, "total_steps": 3752, "loss": 0.4023, "lr": 8.693158039845851e-07, "epoch": 6.4085820895522385, "percentage": 91.55, "elapsed_time": "8:12:20", "remaining_time": "0:45:26"} +{"current_steps": 3440, "total_steps": 3752, "loss": 0.4292, "lr": 8.423857857026574e-07, "epoch": 6.417910447761194, "percentage": 91.68, "elapsed_time": "8:12:57", "remaining_time": "0:44:42"} +{"current_steps": 3445, "total_steps": 3752, "loss": 0.415, "lr": 8.158705070337181e-07, "epoch": 6.427238805970149, "percentage": 91.82, "elapsed_time": "8:13:40", "remaining_time": "0:43:59"} +{"current_steps": 3450, "total_steps": 3752, "loss": 0.417, "lr": 7.897705420020863e-07, "epoch": 6.436567164179104, "percentage": 91.95, "elapsed_time": "8:14:24", "remaining_time": "0:43:16"} +{"current_steps": 3455, "total_steps": 3752, "loss": 0.4281, "lr": 7.640864556410311e-07, "epoch": 6.44589552238806, "percentage": 92.08, "elapsed_time": "8:15:09", "remaining_time": "0:42:33"} +{"current_steps": 3460, "total_steps": 3752, "loss": 0.4097, "lr": 7.388188039805455e-07, "epoch": 6.455223880597015, "percentage": 92.22, "elapsed_time": "8:15:56", "remaining_time": "0:41:51"} +{"current_steps": 3465, "total_steps": 3752, "loss": 0.4036, "lr": 7.139681340352966e-07, "epoch": 6.46455223880597, "percentage": 92.35, "elapsed_time": "8:16:37", "remaining_time": "0:41:08"} +{"current_steps": 3470, "total_steps": 3752, "loss": 0.4073, "lr": 6.895349837928033e-07, "epoch": 6.473880597014926, "percentage": 92.48, "elapsed_time": "8:17:23", "remaining_time": "0:40:25"} +{"current_steps": 3475, "total_steps": 3752, "loss": 0.436, "lr": 6.655198822017661e-07, "epoch": 6.483208955223881, "percentage": 92.62, "elapsed_time": "8:18:07", "remaining_time": "0:39:42"} +{"current_steps": 3480, "total_steps": 3752, "loss": 0.4271, "lr": 6.419233491606403e-07, "epoch": 6.492537313432836, "percentage": 92.75, "elapsed_time": "8:18:53", "remaining_time": "0:38:59"} +{"current_steps": 3485, "total_steps": 3752, "loss": 0.3898, "lr": 6.187458955063652e-07, "epoch": 6.5018656716417915, "percentage": 92.88, "elapsed_time": "8:19:31", "remaining_time": "0:38:16"} +{"current_steps": 3490, "total_steps": 3752, "loss": 0.4292, "lr": 5.959880230033043e-07, "epoch": 6.5111940298507465, "percentage": 93.02, "elapsed_time": "8:20:08", "remaining_time": "0:37:32"} +{"current_steps": 3495, "total_steps": 3752, "loss": 0.4094, "lr": 5.736502243323894e-07, "epoch": 6.520522388059701, "percentage": 93.15, "elapsed_time": "8:20:52", "remaining_time": "0:36:49"} +{"current_steps": 3500, "total_steps": 3752, "loss": 0.3866, "lr": 5.517329830804552e-07, "epoch": 6.529850746268656, "percentage": 93.28, "elapsed_time": "8:21:44", "remaining_time": "0:36:07"} +{"current_steps": 3505, "total_steps": 3752, "loss": 0.3839, "lr": 5.302367737297642e-07, "epoch": 6.539179104477612, "percentage": 93.42, "elapsed_time": "8:22:32", "remaining_time": "0:35:24"} +{"current_steps": 3510, "total_steps": 3752, "loss": 0.4167, "lr": 5.091620616477366e-07, "epoch": 6.548507462686567, "percentage": 93.55, "elapsed_time": "8:23:22", "remaining_time": "0:34:42"} +{"current_steps": 3515, "total_steps": 3752, "loss": 0.3928, "lr": 4.88509303076885e-07, "epoch": 6.557835820895522, "percentage": 93.68, "elapsed_time": "8:24:04", "remaining_time": "0:33:59"} +{"current_steps": 3520, "total_steps": 3752, "loss": 0.4061, "lr": 4.6827894512491814e-07, "epoch": 6.567164179104478, "percentage": 93.82, "elapsed_time": "8:24:51", "remaining_time": "0:33:16"} +{"current_steps": 3525, "total_steps": 3752, "loss": 0.395, "lr": 4.4847142575507706e-07, "epoch": 6.576492537313433, "percentage": 93.95, "elapsed_time": "8:25:35", "remaining_time": "0:32:33"} +{"current_steps": 3530, "total_steps": 3752, "loss": 0.4427, "lr": 4.290871737766544e-07, "epoch": 6.585820895522388, "percentage": 94.08, "elapsed_time": "8:26:07", "remaining_time": "0:31:49"} +{"current_steps": 3535, "total_steps": 3752, "loss": 0.4219, "lr": 4.1012660883570146e-07, "epoch": 6.595149253731344, "percentage": 94.22, "elapsed_time": "8:26:48", "remaining_time": "0:31:06"} +{"current_steps": 3540, "total_steps": 3752, "loss": 0.3875, "lr": 3.915901414059464e-07, "epoch": 6.604477611940299, "percentage": 94.35, "elapsed_time": "8:27:32", "remaining_time": "0:30:23"} +{"current_steps": 3545, "total_steps": 3752, "loss": 0.389, "lr": 3.734781727799086e-07, "epoch": 6.6138059701492535, "percentage": 94.48, "elapsed_time": "8:28:18", "remaining_time": "0:29:40"} +{"current_steps": 3550, "total_steps": 3752, "loss": 0.4182, "lr": 3.557910950602228e-07, "epoch": 6.6231343283582085, "percentage": 94.62, "elapsed_time": "8:29:09", "remaining_time": "0:28:58"} +{"current_steps": 3555, "total_steps": 3752, "loss": 0.3998, "lr": 3.3852929115112177e-07, "epoch": 6.632462686567164, "percentage": 94.75, "elapsed_time": "8:29:50", "remaining_time": "0:28:15"} +{"current_steps": 3560, "total_steps": 3752, "loss": 0.4442, "lr": 3.2169313475018506e-07, "epoch": 6.641791044776119, "percentage": 94.88, "elapsed_time": "8:30:34", "remaining_time": "0:27:32"} +{"current_steps": 3565, "total_steps": 3752, "loss": 0.397, "lr": 3.0528299034021437e-07, "epoch": 6.651119402985074, "percentage": 95.02, "elapsed_time": "8:31:21", "remaining_time": "0:26:49"} +{"current_steps": 3570, "total_steps": 3752, "loss": 0.4202, "lr": 2.892992131813621e-07, "epoch": 6.66044776119403, "percentage": 95.15, "elapsed_time": "8:32:07", "remaining_time": "0:26:06"} +{"current_steps": 3575, "total_steps": 3752, "loss": 0.4157, "lr": 2.737421493034331e-07, "epoch": 6.669776119402985, "percentage": 95.28, "elapsed_time": "8:32:45", "remaining_time": "0:25:23"} +{"current_steps": 3580, "total_steps": 3752, "loss": 0.4163, "lr": 2.586121354984017e-07, "epoch": 6.67910447761194, "percentage": 95.42, "elapsed_time": "8:33:28", "remaining_time": "0:24:40"} +{"current_steps": 3585, "total_steps": 3752, "loss": 0.3958, "lr": 2.4390949931311304e-07, "epoch": 6.688432835820896, "percentage": 95.55, "elapsed_time": "8:34:06", "remaining_time": "0:23:56"} +{"current_steps": 3590, "total_steps": 3752, "loss": 0.4321, "lr": 2.296345590421889e-07, "epoch": 6.697761194029851, "percentage": 95.68, "elapsed_time": "8:34:44", "remaining_time": "0:23:13"} +{"current_steps": 3595, "total_steps": 3752, "loss": 0.3964, "lr": 2.157876237211487e-07, "epoch": 6.707089552238806, "percentage": 95.82, "elapsed_time": "8:35:30", "remaining_time": "0:22:30"} +{"current_steps": 3600, "total_steps": 3752, "loss": 0.4094, "lr": 2.0236899311970593e-07, "epoch": 6.7164179104477615, "percentage": 95.95, "elapsed_time": "8:36:19", "remaining_time": "0:21:48"} +{"current_steps": 3605, "total_steps": 3752, "loss": 0.4351, "lr": 1.8937895773529114e-07, "epoch": 6.725746268656716, "percentage": 96.08, "elapsed_time": "8:37:52", "remaining_time": "0:21:07"} +{"current_steps": 3610, "total_steps": 3752, "loss": 0.4176, "lr": 1.7681779878675697e-07, "epoch": 6.735074626865671, "percentage": 96.22, "elapsed_time": "8:38:29", "remaining_time": "0:20:23"} +{"current_steps": 3615, "total_steps": 3752, "loss": 0.4092, "lr": 1.646857882082853e-07, "epoch": 6.744402985074627, "percentage": 96.35, "elapsed_time": "8:39:11", "remaining_time": "0:19:40"} +{"current_steps": 3620, "total_steps": 3752, "loss": 0.3945, "lr": 1.5298318864350738e-07, "epoch": 6.753731343283582, "percentage": 96.48, "elapsed_time": "8:39:54", "remaining_time": "0:18:57"} +{"current_steps": 3625, "total_steps": 3752, "loss": 0.4233, "lr": 1.417102534398196e-07, "epoch": 6.763059701492537, "percentage": 96.62, "elapsed_time": "8:40:42", "remaining_time": "0:18:14"} +{"current_steps": 3630, "total_steps": 3752, "loss": 0.4228, "lr": 1.3086722664288787e-07, "epoch": 6.772388059701493, "percentage": 96.75, "elapsed_time": "8:41:29", "remaining_time": "0:17:31"} +{"current_steps": 3635, "total_steps": 3752, "loss": 0.405, "lr": 1.2045434299137403e-07, "epoch": 6.781716417910448, "percentage": 96.88, "elapsed_time": "8:42:18", "remaining_time": "0:16:48"} +{"current_steps": 3640, "total_steps": 3752, "loss": 0.4166, "lr": 1.1047182791185551e-07, "epoch": 6.791044776119403, "percentage": 97.01, "elapsed_time": "8:43:01", "remaining_time": "0:16:05"} +{"current_steps": 3645, "total_steps": 3752, "loss": 0.4016, "lr": 1.0091989751393583e-07, "epoch": 6.800373134328359, "percentage": 97.15, "elapsed_time": "8:43:43", "remaining_time": "0:15:22"} +{"current_steps": 3650, "total_steps": 3752, "loss": 0.4096, "lr": 9.179875858557285e-08, "epoch": 6.809701492537314, "percentage": 97.28, "elapsed_time": "8:44:25", "remaining_time": "0:14:39"} +{"current_steps": 3655, "total_steps": 3752, "loss": 0.3898, "lr": 8.310860858859793e-08, "epoch": 6.8190298507462686, "percentage": 97.41, "elapsed_time": "8:45:14", "remaining_time": "0:13:56"} +{"current_steps": 3660, "total_steps": 3752, "loss": 0.4027, "lr": 7.484963565444813e-08, "epoch": 6.8283582089552235, "percentage": 97.55, "elapsed_time": "8:46:05", "remaining_time": "0:13:13"} +{"current_steps": 3665, "total_steps": 3752, "loss": 0.4045, "lr": 6.702201858008517e-08, "epoch": 6.837686567164179, "percentage": 97.68, "elapsed_time": "8:46:43", "remaining_time": "0:12:30"} +{"current_steps": 3670, "total_steps": 3752, "loss": 0.4256, "lr": 5.962592682412948e-08, "epoch": 6.847014925373134, "percentage": 97.81, "elapsed_time": "8:47:34", "remaining_time": "0:11:47"} +{"current_steps": 3675, "total_steps": 3752, "loss": 0.4592, "lr": 5.2661520503192176e-08, "epoch": 6.856343283582089, "percentage": 97.95, "elapsed_time": "8:48:13", "remaining_time": "0:11:04"} +{"current_steps": 3680, "total_steps": 3752, "loss": 0.4153, "lr": 4.612895038840215e-08, "epoch": 6.865671641791045, "percentage": 98.08, "elapsed_time": "8:48:57", "remaining_time": "0:10:20"} +{"current_steps": 3685, "total_steps": 3752, "loss": 0.4372, "lr": 4.0028357902153204e-08, "epoch": 6.875, "percentage": 98.21, "elapsed_time": "8:49:37", "remaining_time": "0:09:37"} +{"current_steps": 3690, "total_steps": 3752, "loss": 0.4054, "lr": 3.435987511503314e-08, "epoch": 6.884328358208955, "percentage": 98.35, "elapsed_time": "8:50:18", "remaining_time": "0:08:54"} +{"current_steps": 3695, "total_steps": 3752, "loss": 0.4037, "lr": 2.9123624742963818e-08, "epoch": 6.893656716417911, "percentage": 98.48, "elapsed_time": "8:51:07", "remaining_time": "0:08:11"} +{"current_steps": 3700, "total_steps": 3752, "loss": 0.4379, "lr": 2.4319720144554417e-08, "epoch": 6.902985074626866, "percentage": 98.61, "elapsed_time": "8:51:50", "remaining_time": "0:07:28"} +{"current_steps": 3705, "total_steps": 3752, "loss": 0.4209, "lr": 1.9948265318638915e-08, "epoch": 6.912313432835821, "percentage": 98.75, "elapsed_time": "8:52:36", "remaining_time": "0:06:45"} +{"current_steps": 3710, "total_steps": 3752, "loss": 0.4224, "lr": 1.6009354902024598e-08, "epoch": 6.9216417910447765, "percentage": 98.88, "elapsed_time": "8:53:22", "remaining_time": "0:06:02"} +{"current_steps": 3715, "total_steps": 3752, "loss": 0.4284, "lr": 1.2503074167451445e-08, "epoch": 6.9309701492537314, "percentage": 99.01, "elapsed_time": "8:54:07", "remaining_time": "0:05:19"} +{"current_steps": 3720, "total_steps": 3752, "loss": 0.4175, "lr": 9.429499021731403e-09, "epoch": 6.940298507462686, "percentage": 99.15, "elapsed_time": "8:54:54", "remaining_time": "0:04:36"} +{"current_steps": 3725, "total_steps": 3752, "loss": 0.4521, "lr": 6.788696004120798e-09, "epoch": 6.949626865671641, "percentage": 99.28, "elapsed_time": "8:55:32", "remaining_time": "0:03:52"} +{"current_steps": 3730, "total_steps": 3752, "loss": 0.3958, "lr": 4.580722284872608e-09, "epoch": 6.958955223880597, "percentage": 99.41, "elapsed_time": "8:56:15", "remaining_time": "0:03:09"} +{"current_steps": 3735, "total_steps": 3752, "loss": 0.3764, "lr": 2.8056256639974467e-09, "epoch": 6.968283582089552, "percentage": 99.55, "elapsed_time": "8:56:57", "remaining_time": "0:02:26"} +{"current_steps": 3740, "total_steps": 3752, "loss": 0.4403, "lr": 1.4634445702288446e-09, "epoch": 6.977611940298507, "percentage": 99.68, "elapsed_time": "8:57:34", "remaining_time": "0:01:43"} +{"current_steps": 3745, "total_steps": 3752, "loss": 0.3982, "lr": 5.542080601950161e-10, "epoch": 6.986940298507463, "percentage": 99.81, "elapsed_time": "8:58:15", "remaining_time": "0:01:00"} +{"current_steps": 3750, "total_steps": 3752, "loss": 0.4035, "lr": 7.793581778825499e-11, "epoch": 6.996268656716418, "percentage": 99.95, "elapsed_time": "8:59:03", "remaining_time": "0:00:17"} +{"current_steps": 3752, "total_steps": 3752, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "9:00:07", "remaining_time": "0:00:00"} diff --git a/trainer_state.json b/trainer_state.json new file mode 100644 index 0000000..bd88f7d --- /dev/null +++ b/trainer_state.json @@ -0,0 +1,8297 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 7.0, + "eval_steps": 500, + "global_step": 3752, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.009328358208955223, + "grad_norm": 10.221339384862345, + "learning_rate": 4.2553191489361704e-07, + "loss": 0.9659, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.47939640283584595, + "step": 5, + "valid_targets_mean": 3400.1, + "valid_targets_min": 951 + }, + { + "epoch": 0.018656716417910446, + "grad_norm": 6.709229009900056, + "learning_rate": 9.574468085106384e-07, + "loss": 0.9484, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40823960304260254, + "step": 10, + "valid_targets_mean": 3299.6, + "valid_targets_min": 1534 + }, + { + "epoch": 0.027985074626865673, + "grad_norm": 6.724610478720323, + "learning_rate": 1.4893617021276596e-06, + "loss": 0.9448, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5101609230041504, + "step": 15, + "valid_targets_mean": 3591.4, + "valid_targets_min": 931 + }, + { + "epoch": 0.03731343283582089, + "grad_norm": 5.060946125149695, + "learning_rate": 2.021276595744681e-06, + "loss": 0.9586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5289248824119568, + "step": 20, + "valid_targets_mean": 4944.5, + "valid_targets_min": 1874 + }, + { + "epoch": 0.04664179104477612, + "grad_norm": 4.365659868837384, + "learning_rate": 2.553191489361702e-06, + "loss": 0.9296, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3888522982597351, + "step": 25, + "valid_targets_mean": 2652.0, + "valid_targets_min": 946 + }, + { + "epoch": 0.055970149253731345, + "grad_norm": 2.7688996011575755, + "learning_rate": 3.0851063829787237e-06, + "loss": 0.8969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29704052209854126, + "step": 30, + "valid_targets_mean": 2535.2, + "valid_targets_min": 714 + }, + { + "epoch": 0.06529850746268656, + "grad_norm": 2.5687899472206235, + "learning_rate": 3.6170212765957453e-06, + "loss": 0.8351, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3777778744697571, + "step": 35, + "valid_targets_mean": 3355.0, + "valid_targets_min": 631 + }, + { + "epoch": 0.07462686567164178, + "grad_norm": 1.7770740261888007, + "learning_rate": 4.148936170212766e-06, + "loss": 0.8389, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4726404845714569, + "step": 40, + "valid_targets_mean": 4328.9, + "valid_targets_min": 831 + }, + { + "epoch": 0.08395522388059702, + "grad_norm": 1.501451727451932, + "learning_rate": 4.680851063829788e-06, + "loss": 0.8262, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.44904133677482605, + "step": 45, + "valid_targets_mean": 3470.4, + "valid_targets_min": 1163 + }, + { + "epoch": 0.09328358208955224, + "grad_norm": 1.044718750816031, + "learning_rate": 5.212765957446809e-06, + "loss": 0.7966, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3058841824531555, + "step": 50, + "valid_targets_mean": 3859.5, + "valid_targets_min": 1479 + }, + { + "epoch": 0.10261194029850747, + "grad_norm": 1.0608825136118278, + "learning_rate": 5.744680851063831e-06, + "loss": 0.8038, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5039680004119873, + "step": 55, + "valid_targets_mean": 4537.5, + "valid_targets_min": 979 + }, + { + "epoch": 0.11194029850746269, + "grad_norm": 1.1185260578222758, + "learning_rate": 6.276595744680851e-06, + "loss": 0.7915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4186343252658844, + "step": 60, + "valid_targets_mean": 2571.9, + "valid_targets_min": 1306 + }, + { + "epoch": 0.12126865671641791, + "grad_norm": 0.9583954658407878, + "learning_rate": 6.808510638297873e-06, + "loss": 0.7856, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3654545545578003, + "step": 65, + "valid_targets_mean": 2627.2, + "valid_targets_min": 773 + }, + { + "epoch": 0.13059701492537312, + "grad_norm": 0.7529579330082371, + "learning_rate": 7.340425531914894e-06, + "loss": 0.7586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24852041900157928, + "step": 70, + "valid_targets_mean": 3186.4, + "valid_targets_min": 654 + }, + { + "epoch": 0.13992537313432835, + "grad_norm": 0.7059756825147047, + "learning_rate": 7.872340425531916e-06, + "loss": 0.7278, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5146458148956299, + "step": 75, + "valid_targets_mean": 7512.9, + "valid_targets_min": 1769 + }, + { + "epoch": 0.14925373134328357, + "grad_norm": 0.840948132821369, + "learning_rate": 8.404255319148937e-06, + "loss": 0.6994, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.323080837726593, + "step": 80, + "valid_targets_mean": 2448.9, + "valid_targets_min": 1430 + }, + { + "epoch": 0.15858208955223882, + "grad_norm": 0.7845909360077343, + "learning_rate": 8.936170212765958e-06, + "loss": 0.7249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31650876998901367, + "step": 85, + "valid_targets_mean": 2846.2, + "valid_targets_min": 296 + }, + { + "epoch": 0.16791044776119404, + "grad_norm": 0.6445084797369868, + "learning_rate": 9.46808510638298e-06, + "loss": 0.6834, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2897588610649109, + "step": 90, + "valid_targets_mean": 4106.8, + "valid_targets_min": 1692 + }, + { + "epoch": 0.17723880597014927, + "grad_norm": 0.870860271288895, + "learning_rate": 1e-05, + "loss": 0.7944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34366244077682495, + "step": 95, + "valid_targets_mean": 3010.6, + "valid_targets_min": 1432 + }, + { + "epoch": 0.1865671641791045, + "grad_norm": 0.750550423240484, + "learning_rate": 1.0531914893617022e-05, + "loss": 0.7016, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3540579676628113, + "step": 100, + "valid_targets_mean": 3653.8, + "valid_targets_min": 2043 + }, + { + "epoch": 0.1958955223880597, + "grad_norm": 0.7284684310754228, + "learning_rate": 1.1063829787234044e-05, + "loss": 0.7004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5650478601455688, + "step": 105, + "valid_targets_mean": 5175.5, + "valid_targets_min": 1609 + }, + { + "epoch": 0.20522388059701493, + "grad_norm": 0.8255701999975512, + "learning_rate": 1.1595744680851065e-05, + "loss": 0.6832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3474902808666229, + "step": 110, + "valid_targets_mean": 3311.2, + "valid_targets_min": 570 + }, + { + "epoch": 0.21455223880597016, + "grad_norm": 0.6481528361446925, + "learning_rate": 1.2127659574468087e-05, + "loss": 0.6627, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2821824252605438, + "step": 115, + "valid_targets_mean": 3884.6, + "valid_targets_min": 1800 + }, + { + "epoch": 0.22388059701492538, + "grad_norm": 0.7670172986923471, + "learning_rate": 1.2659574468085108e-05, + "loss": 0.706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23891422152519226, + "step": 120, + "valid_targets_mean": 2361.6, + "valid_targets_min": 1186 + }, + { + "epoch": 0.2332089552238806, + "grad_norm": 0.6502379252880306, + "learning_rate": 1.3191489361702127e-05, + "loss": 0.6307, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3170034885406494, + "step": 125, + "valid_targets_mean": 4628.2, + "valid_targets_min": 782 + }, + { + "epoch": 0.24253731343283583, + "grad_norm": 0.6975396504222513, + "learning_rate": 1.372340425531915e-05, + "loss": 0.6731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3533723056316376, + "step": 130, + "valid_targets_mean": 3500.1, + "valid_targets_min": 667 + }, + { + "epoch": 0.251865671641791, + "grad_norm": 0.8081018254693072, + "learning_rate": 1.425531914893617e-05, + "loss": 0.6576, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2877732515335083, + "step": 135, + "valid_targets_mean": 2329.9, + "valid_targets_min": 535 + }, + { + "epoch": 0.26119402985074625, + "grad_norm": 0.6874456474101089, + "learning_rate": 1.4787234042553193e-05, + "loss": 0.6502, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37661898136138916, + "step": 140, + "valid_targets_mean": 4439.1, + "valid_targets_min": 1764 + }, + { + "epoch": 0.27052238805970147, + "grad_norm": 0.7322494381218353, + "learning_rate": 1.5319148936170214e-05, + "loss": 0.693, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4837963581085205, + "step": 145, + "valid_targets_mean": 4097.0, + "valid_targets_min": 1553 + }, + { + "epoch": 0.2798507462686567, + "grad_norm": 0.7467656258533556, + "learning_rate": 1.5851063829787235e-05, + "loss": 0.6717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23521462082862854, + "step": 150, + "valid_targets_mean": 2657.5, + "valid_targets_min": 732 + }, + { + "epoch": 0.2891791044776119, + "grad_norm": 0.6821053105109924, + "learning_rate": 1.6382978723404255e-05, + "loss": 0.6199, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30615150928497314, + "step": 155, + "valid_targets_mean": 4557.6, + "valid_targets_min": 1232 + }, + { + "epoch": 0.29850746268656714, + "grad_norm": 0.667618307024502, + "learning_rate": 1.6914893617021276e-05, + "loss": 0.6311, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2656762897968292, + "step": 160, + "valid_targets_mean": 3018.0, + "valid_targets_min": 1023 + }, + { + "epoch": 0.30783582089552236, + "grad_norm": 0.648064122452478, + "learning_rate": 1.74468085106383e-05, + "loss": 0.6028, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28704333305358887, + "step": 165, + "valid_targets_mean": 4269.0, + "valid_targets_min": 1334 + }, + { + "epoch": 0.31716417910447764, + "grad_norm": 0.6741884239379005, + "learning_rate": 1.797872340425532e-05, + "loss": 0.626, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24746419489383698, + "step": 170, + "valid_targets_mean": 2504.9, + "valid_targets_min": 577 + }, + { + "epoch": 0.32649253731343286, + "grad_norm": 0.7022728885755877, + "learning_rate": 1.8510638297872342e-05, + "loss": 0.663, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3848950266838074, + "step": 175, + "valid_targets_mean": 4590.5, + "valid_targets_min": 2201 + }, + { + "epoch": 0.3358208955223881, + "grad_norm": 0.7624122242823003, + "learning_rate": 1.9042553191489363e-05, + "loss": 0.5866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30064308643341064, + "step": 180, + "valid_targets_mean": 3080.8, + "valid_targets_min": 1165 + }, + { + "epoch": 0.3451492537313433, + "grad_norm": 0.8118167793378069, + "learning_rate": 1.9574468085106384e-05, + "loss": 0.6381, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3520858883857727, + "step": 185, + "valid_targets_mean": 2926.2, + "valid_targets_min": 754 + }, + { + "epoch": 0.35447761194029853, + "grad_norm": 0.6793896859604313, + "learning_rate": 2.0106382978723404e-05, + "loss": 0.6347, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.35316750407218933, + "step": 190, + "valid_targets_mean": 3683.0, + "valid_targets_min": 777 + }, + { + "epoch": 0.36380597014925375, + "grad_norm": 0.7198942043975357, + "learning_rate": 2.063829787234043e-05, + "loss": 0.6658, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28977471590042114, + "step": 195, + "valid_targets_mean": 3685.5, + "valid_targets_min": 959 + }, + { + "epoch": 0.373134328358209, + "grad_norm": 0.7721004760482968, + "learning_rate": 2.117021276595745e-05, + "loss": 0.6138, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3721447288990021, + "step": 200, + "valid_targets_mean": 4292.6, + "valid_targets_min": 1399 + }, + { + "epoch": 0.3824626865671642, + "grad_norm": 0.6611030916873262, + "learning_rate": 2.1702127659574467e-05, + "loss": 0.5572, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28156691789627075, + "step": 205, + "valid_targets_mean": 3757.5, + "valid_targets_min": 1029 + }, + { + "epoch": 0.3917910447761194, + "grad_norm": 0.8361909874881175, + "learning_rate": 2.223404255319149e-05, + "loss": 0.6368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2955794036388397, + "step": 210, + "valid_targets_mean": 3336.0, + "valid_targets_min": 1029 + }, + { + "epoch": 0.40111940298507465, + "grad_norm": 0.6111046783642486, + "learning_rate": 2.2765957446808512e-05, + "loss": 0.6236, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24113985896110535, + "step": 215, + "valid_targets_mean": 3413.1, + "valid_targets_min": 1163 + }, + { + "epoch": 0.41044776119402987, + "grad_norm": 0.6936406388230895, + "learning_rate": 2.3297872340425536e-05, + "loss": 0.5931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3664717376232147, + "step": 220, + "valid_targets_mean": 3837.2, + "valid_targets_min": 1589 + }, + { + "epoch": 0.4197761194029851, + "grad_norm": 0.8570856801017971, + "learning_rate": 2.3829787234042553e-05, + "loss": 0.5947, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3282391428947449, + "step": 225, + "valid_targets_mean": 3130.1, + "valid_targets_min": 2013 + }, + { + "epoch": 0.4291044776119403, + "grad_norm": 0.7578881892947891, + "learning_rate": 2.4361702127659578e-05, + "loss": 0.6226, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3216251730918884, + "step": 230, + "valid_targets_mean": 3226.6, + "valid_targets_min": 2393 + }, + { + "epoch": 0.43843283582089554, + "grad_norm": 0.6841426383468427, + "learning_rate": 2.48936170212766e-05, + "loss": 0.5873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29395031929016113, + "step": 235, + "valid_targets_mean": 3245.6, + "valid_targets_min": 777 + }, + { + "epoch": 0.44776119402985076, + "grad_norm": 0.6997097985266442, + "learning_rate": 2.5425531914893616e-05, + "loss": 0.6136, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.278739869594574, + "step": 240, + "valid_targets_mean": 2813.5, + "valid_targets_min": 1595 + }, + { + "epoch": 0.457089552238806, + "grad_norm": 0.8867252786866131, + "learning_rate": 2.595744680851064e-05, + "loss": 0.6212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40658140182495117, + "step": 245, + "valid_targets_mean": 3273.8, + "valid_targets_min": 721 + }, + { + "epoch": 0.4664179104477612, + "grad_norm": 0.7338188790814862, + "learning_rate": 2.6489361702127664e-05, + "loss": 0.5631, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2935730516910553, + "step": 250, + "valid_targets_mean": 3626.6, + "valid_targets_min": 1099 + }, + { + "epoch": 0.47574626865671643, + "grad_norm": 0.7806552708559568, + "learning_rate": 2.702127659574468e-05, + "loss": 0.5797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29288357496261597, + "step": 255, + "valid_targets_mean": 2752.8, + "valid_targets_min": 444 + }, + { + "epoch": 0.48507462686567165, + "grad_norm": 0.7690924986485759, + "learning_rate": 2.7553191489361702e-05, + "loss": 0.5974, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30859023332595825, + "step": 260, + "valid_targets_mean": 2957.0, + "valid_targets_min": 622 + }, + { + "epoch": 0.4944029850746269, + "grad_norm": 0.6851571205475807, + "learning_rate": 2.8085106382978727e-05, + "loss": 0.5887, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2577947676181793, + "step": 265, + "valid_targets_mean": 2963.9, + "valid_targets_min": 961 + }, + { + "epoch": 0.503731343283582, + "grad_norm": 0.7706163388766762, + "learning_rate": 2.8617021276595747e-05, + "loss": 0.6153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2567642033100128, + "step": 270, + "valid_targets_mean": 2558.1, + "valid_targets_min": 913 + }, + { + "epoch": 0.5130597014925373, + "grad_norm": 0.8058419544936287, + "learning_rate": 2.9148936170212768e-05, + "loss": 0.6124, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22249895334243774, + "step": 275, + "valid_targets_mean": 2716.9, + "valid_targets_min": 1002 + }, + { + "epoch": 0.5223880597014925, + "grad_norm": 0.7928536797716753, + "learning_rate": 2.968085106382979e-05, + "loss": 0.6264, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28705018758773804, + "step": 280, + "valid_targets_mean": 2697.9, + "valid_targets_min": 1323 + }, + { + "epoch": 0.5317164179104478, + "grad_norm": 0.7459661026206802, + "learning_rate": 3.0212765957446813e-05, + "loss": 0.6174, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3814293146133423, + "step": 285, + "valid_targets_mean": 4556.2, + "valid_targets_min": 273 + }, + { + "epoch": 0.5410447761194029, + "grad_norm": 0.6769207843617536, + "learning_rate": 3.074468085106383e-05, + "loss": 0.6031, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2440159171819687, + "step": 290, + "valid_targets_mean": 3617.2, + "valid_targets_min": 526 + }, + { + "epoch": 0.5503731343283582, + "grad_norm": 0.7193203456536877, + "learning_rate": 3.127659574468085e-05, + "loss": 0.5742, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23086252808570862, + "step": 295, + "valid_targets_mean": 2440.0, + "valid_targets_min": 491 + }, + { + "epoch": 0.5597014925373134, + "grad_norm": 0.7308475212557202, + "learning_rate": 3.180851063829788e-05, + "loss": 0.6248, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37845319509506226, + "step": 300, + "valid_targets_mean": 3605.0, + "valid_targets_min": 1020 + }, + { + "epoch": 0.5690298507462687, + "grad_norm": 0.9467304828292679, + "learning_rate": 3.234042553191489e-05, + "loss": 0.5664, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2591010630130768, + "step": 305, + "valid_targets_mean": 2529.8, + "valid_targets_min": 1506 + }, + { + "epoch": 0.5783582089552238, + "grad_norm": 0.6697816177321784, + "learning_rate": 3.2872340425531914e-05, + "loss": 0.6137, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23819687962532043, + "step": 310, + "valid_targets_mean": 3026.1, + "valid_targets_min": 257 + }, + { + "epoch": 0.5876865671641791, + "grad_norm": 0.6975396328455103, + "learning_rate": 3.340425531914894e-05, + "loss": 0.6167, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20802253484725952, + "step": 315, + "valid_targets_mean": 2358.5, + "valid_targets_min": 1528 + }, + { + "epoch": 0.5970149253731343, + "grad_norm": 0.7175195462660079, + "learning_rate": 3.393617021276596e-05, + "loss": 0.6265, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42556458711624146, + "step": 320, + "valid_targets_mean": 5309.6, + "valid_targets_min": 1366 + }, + { + "epoch": 0.6063432835820896, + "grad_norm": 0.681879973834109, + "learning_rate": 3.446808510638298e-05, + "loss": 0.5727, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3077801764011383, + "step": 325, + "valid_targets_mean": 4259.2, + "valid_targets_min": 2006 + }, + { + "epoch": 0.6156716417910447, + "grad_norm": 0.7422054132999161, + "learning_rate": 3.5000000000000004e-05, + "loss": 0.6388, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2498745322227478, + "step": 330, + "valid_targets_mean": 3084.8, + "valid_targets_min": 1479 + }, + { + "epoch": 0.625, + "grad_norm": 0.7816443491468698, + "learning_rate": 3.5531914893617025e-05, + "loss": 0.5431, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17738372087478638, + "step": 335, + "valid_targets_mean": 1923.5, + "valid_targets_min": 392 + }, + { + "epoch": 0.6343283582089553, + "grad_norm": 0.6926315605053619, + "learning_rate": 3.6063829787234045e-05, + "loss": 0.5734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1904895007610321, + "step": 340, + "valid_targets_mean": 2476.5, + "valid_targets_min": 840 + }, + { + "epoch": 0.6436567164179104, + "grad_norm": 0.7927663242744208, + "learning_rate": 3.6595744680851066e-05, + "loss": 0.5895, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2929850220680237, + "step": 345, + "valid_targets_mean": 2753.6, + "valid_targets_min": 920 + }, + { + "epoch": 0.6529850746268657, + "grad_norm": 0.6667734659693237, + "learning_rate": 3.712765957446809e-05, + "loss": 0.5856, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2891857922077179, + "step": 350, + "valid_targets_mean": 4494.9, + "valid_targets_min": 313 + }, + { + "epoch": 0.6623134328358209, + "grad_norm": 0.6373264026859415, + "learning_rate": 3.7659574468085114e-05, + "loss": 0.5908, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2208825796842575, + "step": 355, + "valid_targets_mean": 3450.0, + "valid_targets_min": 1102 + }, + { + "epoch": 0.6716417910447762, + "grad_norm": 0.7594230097038847, + "learning_rate": 3.819148936170213e-05, + "loss": 0.6502, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4106304347515106, + "step": 360, + "valid_targets_mean": 3784.1, + "valid_targets_min": 1085 + }, + { + "epoch": 0.6809701492537313, + "grad_norm": 0.8582961346468819, + "learning_rate": 3.872340425531915e-05, + "loss": 0.5736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4053879380226135, + "step": 365, + "valid_targets_mean": 4927.6, + "valid_targets_min": 385 + }, + { + "epoch": 0.6902985074626866, + "grad_norm": 0.7969092384416857, + "learning_rate": 3.925531914893618e-05, + "loss": 0.5662, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2686024606227875, + "step": 370, + "valid_targets_mean": 2672.2, + "valid_targets_min": 336 + }, + { + "epoch": 0.6996268656716418, + "grad_norm": 0.7347204620089871, + "learning_rate": 3.978723404255319e-05, + "loss": 0.559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21485693752765656, + "step": 375, + "valid_targets_mean": 3369.1, + "valid_targets_min": 1116 + }, + { + "epoch": 0.7089552238805971, + "grad_norm": 0.6164343450129072, + "learning_rate": 3.999992206418221e-05, + "loss": 0.5744, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.38275283575057983, + "step": 380, + "valid_targets_mean": 5458.0, + "valid_targets_min": 1447 + }, + { + "epoch": 0.7182835820895522, + "grad_norm": 0.7884354979827474, + "learning_rate": 3.999944579193981e-05, + "loss": 0.611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25685256719589233, + "step": 385, + "valid_targets_mean": 2556.9, + "valid_targets_min": 770 + }, + { + "epoch": 0.7276119402985075, + "grad_norm": 0.7217636866372948, + "learning_rate": 3.999853655542977e-05, + "loss": 0.5903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2114216685295105, + "step": 390, + "valid_targets_mean": 2227.6, + "valid_targets_min": 1368 + }, + { + "epoch": 0.7369402985074627, + "grad_norm": 0.6244477830330477, + "learning_rate": 3.999719437433601e-05, + "loss": 0.5938, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29096540808677673, + "step": 395, + "valid_targets_mean": 4716.9, + "valid_targets_min": 309 + }, + { + "epoch": 0.746268656716418, + "grad_norm": 0.5648390962498896, + "learning_rate": 3.999541927771513e-05, + "loss": 0.5653, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3506127595901489, + "step": 400, + "valid_targets_mean": 5497.1, + "valid_targets_min": 1810 + }, + { + "epoch": 0.7555970149253731, + "grad_norm": 0.6874932901853528, + "learning_rate": 3.999321130399588e-05, + "loss": 0.5915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3548329770565033, + "step": 405, + "valid_targets_mean": 4386.9, + "valid_targets_min": 1639 + }, + { + "epoch": 0.7649253731343284, + "grad_norm": 0.6695661761975534, + "learning_rate": 3.9990570500978275e-05, + "loss": 0.6022, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24912483990192413, + "step": 410, + "valid_targets_mean": 3129.9, + "valid_targets_min": 1244 + }, + { + "epoch": 0.7742537313432836, + "grad_norm": 0.8829757467321712, + "learning_rate": 3.998749692583255e-05, + "loss": 0.5879, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3693455457687378, + "step": 415, + "valid_targets_mean": 4967.1, + "valid_targets_min": 1224 + }, + { + "epoch": 0.7835820895522388, + "grad_norm": 0.7060180533085941, + "learning_rate": 3.9983990645097977e-05, + "loss": 0.5516, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31799471378326416, + "step": 420, + "valid_targets_mean": 3629.6, + "valid_targets_min": 375 + }, + { + "epoch": 0.792910447761194, + "grad_norm": 0.6512688877556654, + "learning_rate": 3.9980051734681365e-05, + "loss": 0.5906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24405537545681, + "step": 425, + "valid_targets_mean": 3651.1, + "valid_targets_min": 1555 + }, + { + "epoch": 0.8022388059701493, + "grad_norm": 0.7659981510033175, + "learning_rate": 3.9975680279855453e-05, + "loss": 0.571, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3422507047653198, + "step": 430, + "valid_targets_mean": 3973.5, + "valid_targets_min": 2179 + }, + { + "epoch": 0.8115671641791045, + "grad_norm": 0.6393299412177378, + "learning_rate": 3.9970876375257044e-05, + "loss": 0.5607, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24772992730140686, + "step": 435, + "valid_targets_mean": 4269.0, + "valid_targets_min": 1172 + }, + { + "epoch": 0.8208955223880597, + "grad_norm": 0.631301335819845, + "learning_rate": 3.996564012488497e-05, + "loss": 0.58, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34593862295150757, + "step": 440, + "valid_targets_mean": 5033.6, + "valid_targets_min": 2351 + }, + { + "epoch": 0.8302238805970149, + "grad_norm": 0.6978871998832528, + "learning_rate": 3.995997164209785e-05, + "loss": 0.5774, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3239847421646118, + "step": 445, + "valid_targets_mean": 3281.9, + "valid_targets_min": 753 + }, + { + "epoch": 0.8395522388059702, + "grad_norm": 0.7500953271176343, + "learning_rate": 3.99538710496116e-05, + "loss": 0.5685, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3513426184654236, + "step": 450, + "valid_targets_mean": 3056.2, + "valid_targets_min": 1119 + }, + { + "epoch": 0.8488805970149254, + "grad_norm": 0.5213096901949963, + "learning_rate": 3.9947338479496813e-05, + "loss": 0.5592, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2794053256511688, + "step": 455, + "valid_targets_mean": 6154.8, + "valid_targets_min": 736 + }, + { + "epoch": 0.8582089552238806, + "grad_norm": 0.6469270473153445, + "learning_rate": 3.994037407317588e-05, + "loss": 0.5656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3254997730255127, + "step": 460, + "valid_targets_mean": 4166.6, + "valid_targets_min": 2197 + }, + { + "epoch": 0.8675373134328358, + "grad_norm": 0.7218976843566365, + "learning_rate": 3.993297798141992e-05, + "loss": 0.6036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30365434288978577, + "step": 465, + "valid_targets_mean": 3196.2, + "valid_targets_min": 676 + }, + { + "epoch": 0.8768656716417911, + "grad_norm": 0.6248269334152132, + "learning_rate": 3.9925150364345556e-05, + "loss": 0.6171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27022770047187805, + "step": 470, + "valid_targets_mean": 4472.8, + "valid_targets_min": 2442 + }, + { + "epoch": 0.8861940298507462, + "grad_norm": 0.6281545909117507, + "learning_rate": 3.991689139141141e-05, + "loss": 0.5776, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40134453773498535, + "step": 475, + "valid_targets_mean": 4269.2, + "valid_targets_min": 1258 + }, + { + "epoch": 0.8955223880597015, + "grad_norm": 0.6024513638877718, + "learning_rate": 3.9908201241414434e-05, + "loss": 0.5608, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24152323603630066, + "step": 480, + "valid_targets_mean": 3675.8, + "valid_targets_min": 1462 + }, + { + "epoch": 0.9048507462686567, + "grad_norm": 0.6710355134691387, + "learning_rate": 3.9899080102486065e-05, + "loss": 0.5632, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1799512654542923, + "step": 485, + "valid_targets_mean": 2174.6, + "valid_targets_min": 648 + }, + { + "epoch": 0.914179104477612, + "grad_norm": 0.5866994211338196, + "learning_rate": 3.988952817208815e-05, + "loss": 0.5421, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27179545164108276, + "step": 490, + "valid_targets_mean": 4399.2, + "valid_targets_min": 1815 + }, + { + "epoch": 0.9235074626865671, + "grad_norm": 0.5898315810107299, + "learning_rate": 3.987954565700863e-05, + "loss": 0.5646, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3146028518676758, + "step": 495, + "valid_targets_mean": 4055.4, + "valid_targets_min": 1896 + }, + { + "epoch": 0.9328358208955224, + "grad_norm": 0.8912745785022403, + "learning_rate": 3.9869132773357115e-05, + "loss": 0.5626, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3271094262599945, + "step": 500, + "valid_targets_mean": 3561.0, + "valid_targets_min": 1819 + }, + { + "epoch": 0.9421641791044776, + "grad_norm": 0.6881650874244652, + "learning_rate": 3.9858289746560183e-05, + "loss": 0.5613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2683401107788086, + "step": 505, + "valid_targets_mean": 3099.0, + "valid_targets_min": 1497 + }, + { + "epoch": 0.9514925373134329, + "grad_norm": 0.66620144560675, + "learning_rate": 3.98470168113565e-05, + "loss": 0.5653, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25099435448646545, + "step": 510, + "valid_targets_mean": 3040.9, + "valid_targets_min": 790 + }, + { + "epoch": 0.960820895522388, + "grad_norm": 0.7214014750128713, + "learning_rate": 3.983531421179172e-05, + "loss": 0.5751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26236626505851746, + "step": 515, + "valid_targets_mean": 2553.9, + "valid_targets_min": 1640 + }, + { + "epoch": 0.9701492537313433, + "grad_norm": 0.6319763749803543, + "learning_rate": 3.9823182201213245e-05, + "loss": 0.5968, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3791792690753937, + "step": 520, + "valid_targets_mean": 4996.2, + "valid_targets_min": 1588 + }, + { + "epoch": 0.9794776119402985, + "grad_norm": 0.709042434719836, + "learning_rate": 3.981062104226471e-05, + "loss": 0.5553, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3071591854095459, + "step": 525, + "valid_targets_mean": 3725.9, + "valid_targets_min": 1311 + }, + { + "epoch": 0.9888059701492538, + "grad_norm": 0.630910177797118, + "learning_rate": 3.97976310068803e-05, + "loss": 0.6386, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20051100850105286, + "step": 530, + "valid_targets_mean": 3030.5, + "valid_targets_min": 1116 + }, + { + "epoch": 0.9981343283582089, + "grad_norm": 0.7022547389585898, + "learning_rate": 3.978421237627886e-05, + "loss": 0.5233, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27863332629203796, + "step": 535, + "valid_targets_mean": 3701.6, + "valid_targets_min": 1689 + }, + { + "epoch": 1.007462686567164, + "grad_norm": 0.8028854115490012, + "learning_rate": 3.977036544095781e-05, + "loss": 0.553, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2280224859714508, + "step": 540, + "valid_targets_mean": 3242.6, + "valid_targets_min": 746 + }, + { + "epoch": 1.0167910447761195, + "grad_norm": 0.6315346456732163, + "learning_rate": 3.975609050068689e-05, + "loss": 0.5425, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17604753375053406, + "step": 545, + "valid_targets_mean": 2877.9, + "valid_targets_min": 1502 + }, + { + "epoch": 1.0261194029850746, + "grad_norm": 0.5900002807896124, + "learning_rate": 3.97413878645016e-05, + "loss": 0.5363, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19828876852989197, + "step": 550, + "valid_targets_mean": 3338.2, + "valid_targets_min": 1626 + }, + { + "epoch": 1.0354477611940298, + "grad_norm": 0.6785431971677857, + "learning_rate": 3.972625785069657e-05, + "loss": 0.5389, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3047122657299042, + "step": 555, + "valid_targets_mean": 4309.5, + "valid_targets_min": 1341 + }, + { + "epoch": 1.044776119402985, + "grad_norm": 0.6507862460759172, + "learning_rate": 3.971070078681864e-05, + "loss": 0.5043, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22440937161445618, + "step": 560, + "valid_targets_mean": 4261.5, + "valid_targets_min": 1296 + }, + { + "epoch": 1.0541044776119404, + "grad_norm": 0.7460700856333692, + "learning_rate": 3.969471700965979e-05, + "loss": 0.537, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25620323419570923, + "step": 565, + "valid_targets_mean": 2881.9, + "valid_targets_min": 1108 + }, + { + "epoch": 1.0634328358208955, + "grad_norm": 0.7591233299699717, + "learning_rate": 3.967830686524982e-05, + "loss": 0.5388, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2694655656814575, + "step": 570, + "valid_targets_mean": 3030.0, + "valid_targets_min": 1663 + }, + { + "epoch": 1.0727611940298507, + "grad_norm": 0.6643950474255416, + "learning_rate": 3.966147070884888e-05, + "loss": 0.5375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20027458667755127, + "step": 575, + "valid_targets_mean": 4767.1, + "valid_targets_min": 820 + }, + { + "epoch": 1.0820895522388059, + "grad_norm": 0.6045465434953375, + "learning_rate": 3.964420890493978e-05, + "loss": 0.5429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28705334663391113, + "step": 580, + "valid_targets_mean": 5348.8, + "valid_targets_min": 1163 + }, + { + "epoch": 1.0914179104477613, + "grad_norm": 0.6431442021457563, + "learning_rate": 3.9626521827220096e-05, + "loss": 0.5274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25629544258117676, + "step": 585, + "valid_targets_mean": 3699.1, + "valid_targets_min": 1552 + }, + { + "epoch": 1.1007462686567164, + "grad_norm": 0.6892604477150721, + "learning_rate": 3.960840985859406e-05, + "loss": 0.5418, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3467318117618561, + "step": 590, + "valid_targets_mean": 3772.8, + "valid_targets_min": 1645 + }, + { + "epoch": 1.1100746268656716, + "grad_norm": 0.598272640729981, + "learning_rate": 3.95898733911643e-05, + "loss": 0.5166, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4067884087562561, + "step": 595, + "valid_targets_mean": 6477.9, + "valid_targets_min": 1571 + }, + { + "epoch": 1.1194029850746268, + "grad_norm": 0.7278171904310045, + "learning_rate": 3.957091282622335e-05, + "loss": 0.5539, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25010570883750916, + "step": 600, + "valid_targets_mean": 3534.5, + "valid_targets_min": 1070 + }, + { + "epoch": 1.1287313432835822, + "grad_norm": 0.786396397769096, + "learning_rate": 3.955152857424493e-05, + "loss": 0.5375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2975982427597046, + "step": 605, + "valid_targets_mean": 2510.1, + "valid_targets_min": 829 + }, + { + "epoch": 1.1380597014925373, + "grad_norm": 0.685880736946905, + "learning_rate": 3.953172105487509e-05, + "loss": 0.5554, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2779093384742737, + "step": 610, + "valid_targets_mean": 3502.4, + "valid_targets_min": 1914 + }, + { + "epoch": 1.1473880597014925, + "grad_norm": 0.5820074108320361, + "learning_rate": 3.951149069692312e-05, + "loss": 0.5316, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2435397058725357, + "step": 615, + "valid_targets_mean": 4565.1, + "valid_targets_min": 1057 + }, + { + "epoch": 1.1567164179104479, + "grad_norm": 0.6468224388409047, + "learning_rate": 3.9490837938352267e-05, + "loss": 0.5774, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19375991821289062, + "step": 620, + "valid_targets_mean": 3120.6, + "valid_targets_min": 1164 + }, + { + "epoch": 1.166044776119403, + "grad_norm": 0.7850368185324255, + "learning_rate": 3.946976322627024e-05, + "loss": 0.555, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28971314430236816, + "step": 625, + "valid_targets_mean": 2897.0, + "valid_targets_min": 951 + }, + { + "epoch": 1.1753731343283582, + "grad_norm": 0.6912538408752664, + "learning_rate": 3.944826701691955e-05, + "loss": 0.5606, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27505776286125183, + "step": 630, + "valid_targets_mean": 4580.4, + "valid_targets_min": 786 + }, + { + "epoch": 1.1847014925373134, + "grad_norm": 0.7600526449608063, + "learning_rate": 3.942634977566761e-05, + "loss": 0.5389, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24214911460876465, + "step": 635, + "valid_targets_mean": 2262.8, + "valid_targets_min": 732 + }, + { + "epoch": 1.1940298507462686, + "grad_norm": 0.632283784633046, + "learning_rate": 3.9404011976996696e-05, + "loss": 0.6084, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21673955023288727, + "step": 640, + "valid_targets_mean": 3801.0, + "valid_targets_min": 1953 + }, + { + "epoch": 1.203358208955224, + "grad_norm": 0.6289924112629002, + "learning_rate": 3.938125410449364e-05, + "loss": 0.5168, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2503665089607239, + "step": 645, + "valid_targets_mean": 3704.6, + "valid_targets_min": 856 + }, + { + "epoch": 1.212686567164179, + "grad_norm": 0.7344098110742897, + "learning_rate": 3.935807665083936e-05, + "loss": 0.5465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.36894696950912476, + "step": 650, + "valid_targets_mean": 3159.4, + "valid_targets_min": 296 + }, + { + "epoch": 1.2220149253731343, + "grad_norm": 0.5804792478675688, + "learning_rate": 3.933448011779824e-05, + "loss": 0.5323, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15652263164520264, + "step": 655, + "valid_targets_mean": 2600.0, + "valid_targets_min": 1891 + }, + { + "epoch": 1.2313432835820897, + "grad_norm": 0.6322902909680553, + "learning_rate": 3.93104650162072e-05, + "loss": 0.5395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3222752809524536, + "step": 660, + "valid_targets_mean": 4437.5, + "valid_targets_min": 1406 + }, + { + "epoch": 1.2406716417910448, + "grad_norm": 0.6164194377591228, + "learning_rate": 3.928603186596471e-05, + "loss": 0.5113, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2050849050283432, + "step": 665, + "valid_targets_mean": 3560.0, + "valid_targets_min": 877 + }, + { + "epoch": 1.25, + "grad_norm": 0.707935496535295, + "learning_rate": 3.926118119601946e-05, + "loss": 0.5832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28239184617996216, + "step": 670, + "valid_targets_mean": 3334.2, + "valid_targets_min": 1734 + }, + { + "epoch": 1.2593283582089552, + "grad_norm": 0.5849604176220515, + "learning_rate": 3.923591354435897e-05, + "loss": 0.5735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29911866784095764, + "step": 675, + "valid_targets_mean": 4364.2, + "valid_targets_min": 1916 + }, + { + "epoch": 1.2686567164179103, + "grad_norm": 0.6287509000150424, + "learning_rate": 3.9210229457997916e-05, + "loss": 0.5395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2777480483055115, + "step": 680, + "valid_targets_mean": 4163.1, + "valid_targets_min": 1859 + }, + { + "epoch": 1.2779850746268657, + "grad_norm": 0.685288777518089, + "learning_rate": 3.9184129492966286e-05, + "loss": 0.4891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31055453419685364, + "step": 685, + "valid_targets_mean": 3987.6, + "valid_targets_min": 2119 + }, + { + "epoch": 1.287313432835821, + "grad_norm": 0.6339721066974396, + "learning_rate": 3.915761421429735e-05, + "loss": 0.522, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32543087005615234, + "step": 690, + "valid_targets_mean": 4985.5, + "valid_targets_min": 2715 + }, + { + "epoch": 1.296641791044776, + "grad_norm": 0.5733697106249588, + "learning_rate": 3.913068419601542e-05, + "loss": 0.5745, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3147970736026764, + "step": 695, + "valid_targets_mean": 4724.9, + "valid_targets_min": 1953 + }, + { + "epoch": 1.3059701492537314, + "grad_norm": 0.7724230248030342, + "learning_rate": 3.910334002112344e-05, + "loss": 0.5773, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3536461889743805, + "step": 700, + "valid_targets_mean": 3440.6, + "valid_targets_min": 1399 + }, + { + "epoch": 1.3152985074626866, + "grad_norm": 0.8049633644751847, + "learning_rate": 3.907558228159034e-05, + "loss": 0.5229, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24340598285198212, + "step": 705, + "valid_targets_mean": 3403.8, + "valid_targets_min": 1012 + }, + { + "epoch": 1.3246268656716418, + "grad_norm": 0.5473414470233305, + "learning_rate": 3.904741157833825e-05, + "loss": 0.5317, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20663416385650635, + "step": 710, + "valid_targets_mean": 3928.6, + "valid_targets_min": 1371 + }, + { + "epoch": 1.333955223880597, + "grad_norm": 0.7259815989545045, + "learning_rate": 3.901882852122945e-05, + "loss": 0.5331, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2875136733055115, + "step": 715, + "valid_targets_mean": 3111.2, + "valid_targets_min": 354 + }, + { + "epoch": 1.3432835820895521, + "grad_norm": 1.6864100894185075, + "learning_rate": 3.8989833729053226e-05, + "loss": 0.5126, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21058915555477142, + "step": 720, + "valid_targets_mean": 3734.4, + "valid_targets_min": 1357 + }, + { + "epoch": 1.3526119402985075, + "grad_norm": 0.7096608255241813, + "learning_rate": 3.89604278295124e-05, + "loss": 0.5341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25622522830963135, + "step": 725, + "valid_targets_mean": 2885.2, + "valid_targets_min": 1001 + }, + { + "epoch": 1.3619402985074627, + "grad_norm": 0.7945646267043578, + "learning_rate": 3.893061145920981e-05, + "loss": 0.5368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2410912960767746, + "step": 730, + "valid_targets_mean": 4002.9, + "valid_targets_min": 960 + }, + { + "epoch": 1.3712686567164178, + "grad_norm": 0.5704159076561983, + "learning_rate": 3.890038526363449e-05, + "loss": 0.5101, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3112795948982239, + "step": 735, + "valid_targets_mean": 5595.6, + "valid_targets_min": 1909 + }, + { + "epoch": 1.3805970149253732, + "grad_norm": 0.573903643156363, + "learning_rate": 3.88697498971477e-05, + "loss": 0.4972, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19024305045604706, + "step": 740, + "valid_targets_mean": 3188.5, + "valid_targets_min": 901 + }, + { + "epoch": 1.3899253731343284, + "grad_norm": 0.6681165438144722, + "learning_rate": 3.883870602296878e-05, + "loss": 0.4791, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15833920240402222, + "step": 745, + "valid_targets_mean": 2132.2, + "valid_targets_min": 693 + }, + { + "epoch": 1.3992537313432836, + "grad_norm": 0.693907557505896, + "learning_rate": 3.880725431316074e-05, + "loss": 0.5674, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3302950859069824, + "step": 750, + "valid_targets_mean": 3976.1, + "valid_targets_min": 1655 + }, + { + "epoch": 1.4085820895522387, + "grad_norm": 0.6705694993669753, + "learning_rate": 3.87753954486158e-05, + "loss": 0.5356, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17933990061283112, + "step": 755, + "valid_targets_mean": 2045.2, + "valid_targets_min": 512 + }, + { + "epoch": 1.417910447761194, + "grad_norm": 0.5580232360061667, + "learning_rate": 3.874313011904056e-05, + "loss": 0.5496, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37407204508781433, + "step": 760, + "valid_targets_mean": 5818.9, + "valid_targets_min": 1488 + }, + { + "epoch": 1.4272388059701493, + "grad_norm": 0.872403546054947, + "learning_rate": 3.871045902294112e-05, + "loss": 0.5187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.35408133268356323, + "step": 765, + "valid_targets_mean": 3871.9, + "valid_targets_min": 1941 + }, + { + "epoch": 1.4365671641791045, + "grad_norm": 0.7004342757805163, + "learning_rate": 3.867738286760793e-05, + "loss": 0.5382, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30939149856567383, + "step": 770, + "valid_targets_mean": 3215.5, + "valid_targets_min": 2319 + }, + { + "epoch": 1.4458955223880596, + "grad_norm": 0.703984049995771, + "learning_rate": 3.86439023691005e-05, + "loss": 0.518, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32864096760749817, + "step": 775, + "valid_targets_mean": 3294.1, + "valid_targets_min": 673 + }, + { + "epoch": 1.455223880597015, + "grad_norm": 0.7228480175806797, + "learning_rate": 3.86100182522319e-05, + "loss": 0.5527, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22931042313575745, + "step": 780, + "valid_targets_mean": 2563.4, + "valid_targets_min": 820 + }, + { + "epoch": 1.4645522388059702, + "grad_norm": 0.6751507060984445, + "learning_rate": 3.857573125055304e-05, + "loss": 0.5226, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3485998511314392, + "step": 785, + "valid_targets_mean": 3630.6, + "valid_targets_min": 1150 + }, + { + "epoch": 1.4738805970149254, + "grad_norm": 0.6123224789955897, + "learning_rate": 3.8541042106336825e-05, + "loss": 0.548, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22629578411579132, + "step": 790, + "valid_targets_mean": 4140.4, + "valid_targets_min": 987 + }, + { + "epoch": 1.4832089552238805, + "grad_norm": 0.6299059225542767, + "learning_rate": 3.850595157056206e-05, + "loss": 0.5544, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2941248416900635, + "step": 795, + "valid_targets_mean": 3076.1, + "valid_targets_min": 449 + }, + { + "epoch": 1.4925373134328357, + "grad_norm": 0.6400254246581368, + "learning_rate": 3.84704604028972e-05, + "loss": 0.5091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21500684320926666, + "step": 800, + "valid_targets_mean": 3791.8, + "valid_targets_min": 1801 + }, + { + "epoch": 1.501865671641791, + "grad_norm": 0.6978713655582434, + "learning_rate": 3.84345693716839e-05, + "loss": 0.529, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26665663719177246, + "step": 805, + "valid_targets_mean": 3161.4, + "valid_targets_min": 1303 + }, + { + "epoch": 1.5111940298507462, + "grad_norm": 0.6380967850517599, + "learning_rate": 3.8398279253920403e-05, + "loss": 0.5368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27644985914230347, + "step": 810, + "valid_targets_mean": 2991.6, + "valid_targets_min": 1755 + }, + { + "epoch": 1.5205223880597014, + "grad_norm": 0.5852970318086799, + "learning_rate": 3.8361590835244696e-05, + "loss": 0.535, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2506909668445587, + "step": 815, + "valid_targets_mean": 4482.4, + "valid_targets_min": 2066 + }, + { + "epoch": 1.5298507462686568, + "grad_norm": 0.6543499665228387, + "learning_rate": 3.8324504909917486e-05, + "loss": 0.533, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26362523436546326, + "step": 820, + "valid_targets_mean": 3188.2, + "valid_targets_min": 1237 + }, + { + "epoch": 1.539179104477612, + "grad_norm": 0.5330651622885079, + "learning_rate": 3.8287022280805064e-05, + "loss": 0.5049, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1682763248682022, + "step": 825, + "valid_targets_mean": 3644.2, + "valid_targets_min": 1881 + }, + { + "epoch": 1.5485074626865671, + "grad_norm": 0.7168842466101432, + "learning_rate": 3.824914375936186e-05, + "loss": 0.5542, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28634709119796753, + "step": 830, + "valid_targets_mean": 3268.8, + "valid_targets_min": 1515 + }, + { + "epoch": 1.5578358208955225, + "grad_norm": 0.6795250702425825, + "learning_rate": 3.8210870165612913e-05, + "loss": 0.5247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23117414116859436, + "step": 835, + "valid_targets_mean": 4144.9, + "valid_targets_min": 2050 + }, + { + "epoch": 1.5671641791044775, + "grad_norm": 0.5422177330938823, + "learning_rate": 3.817220232813611e-05, + "loss": 0.525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31314602494239807, + "step": 840, + "valid_targets_mean": 6506.9, + "valid_targets_min": 375 + }, + { + "epoch": 1.5764925373134329, + "grad_norm": 0.6420269794456652, + "learning_rate": 3.8133141084044243e-05, + "loss": 0.5285, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3007183074951172, + "step": 845, + "valid_targets_mean": 4209.8, + "valid_targets_min": 1335 + }, + { + "epoch": 1.585820895522388, + "grad_norm": 0.8282572723489329, + "learning_rate": 3.809368727896689e-05, + "loss": 0.5223, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2658495604991913, + "step": 850, + "valid_targets_mean": 2924.5, + "valid_targets_min": 1367 + }, + { + "epoch": 1.5951492537313432, + "grad_norm": 0.5631240642789294, + "learning_rate": 3.805384176703211e-05, + "loss": 0.5296, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34055593609809875, + "step": 855, + "valid_targets_mean": 5657.5, + "valid_targets_min": 2193 + }, + { + "epoch": 1.6044776119402986, + "grad_norm": 0.5750622347506824, + "learning_rate": 3.8013605410847956e-05, + "loss": 0.5587, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2550086975097656, + "step": 860, + "valid_targets_mean": 4503.9, + "valid_targets_min": 1624 + }, + { + "epoch": 1.6138059701492538, + "grad_norm": 0.7262918670089581, + "learning_rate": 3.797297908148379e-05, + "loss": 0.546, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3380868434906006, + "step": 865, + "valid_targets_mean": 3818.6, + "valid_targets_min": 1504 + }, + { + "epoch": 1.623134328358209, + "grad_norm": 0.6021587530457694, + "learning_rate": 3.793196365845142e-05, + "loss": 0.5525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2848277986049652, + "step": 870, + "valid_targets_mean": 4457.8, + "valid_targets_min": 2198 + }, + { + "epoch": 1.6324626865671643, + "grad_norm": 0.7091437572796929, + "learning_rate": 3.789056002968609e-05, + "loss": 0.5657, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2223820686340332, + "step": 875, + "valid_targets_mean": 2859.2, + "valid_targets_min": 916 + }, + { + "epoch": 1.6417910447761193, + "grad_norm": 0.6483019545331663, + "learning_rate": 3.784876909152723e-05, + "loss": 0.4998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17174361646175385, + "step": 880, + "valid_targets_mean": 2351.5, + "valid_targets_min": 1072 + }, + { + "epoch": 1.6511194029850746, + "grad_norm": 0.6542063501596606, + "learning_rate": 3.7806591748699056e-05, + "loss": 0.5349, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2808379828929901, + "step": 885, + "valid_targets_mean": 3549.6, + "valid_targets_min": 1260 + }, + { + "epoch": 1.6604477611940298, + "grad_norm": 0.5937691457660903, + "learning_rate": 3.776402891429098e-05, + "loss": 0.5024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22422434389591217, + "step": 890, + "valid_targets_mean": 3772.2, + "valid_targets_min": 1460 + }, + { + "epoch": 1.669776119402985, + "grad_norm": 0.7015171188301872, + "learning_rate": 3.772108150973786e-05, + "loss": 0.5186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2726173400878906, + "step": 895, + "valid_targets_mean": 2988.9, + "valid_targets_min": 694 + }, + { + "epoch": 1.6791044776119404, + "grad_norm": 0.6708109895007804, + "learning_rate": 3.767775046480004e-05, + "loss": 0.6051, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3386884927749634, + "step": 900, + "valid_targets_mean": 3630.9, + "valid_targets_min": 2197 + }, + { + "epoch": 1.6884328358208955, + "grad_norm": 0.6667077027857159, + "learning_rate": 3.7634036717543224e-05, + "loss": 0.5367, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3334267735481262, + "step": 905, + "valid_targets_mean": 3851.1, + "valid_targets_min": 1222 + }, + { + "epoch": 1.6977611940298507, + "grad_norm": 0.6029798187221745, + "learning_rate": 3.7589941214318156e-05, + "loss": 0.5028, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17591658234596252, + "step": 910, + "valid_targets_mean": 2475.8, + "valid_targets_min": 1297 + }, + { + "epoch": 1.707089552238806, + "grad_norm": 0.6695801867461935, + "learning_rate": 3.754546490974016e-05, + "loss": 0.5176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.38722145557403564, + "step": 915, + "valid_targets_mean": 3990.0, + "valid_targets_min": 1180 + }, + { + "epoch": 1.716417910447761, + "grad_norm": 0.6945458634583573, + "learning_rate": 3.750060876666844e-05, + "loss": 0.5256, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32125023007392883, + "step": 920, + "valid_targets_mean": 3428.8, + "valid_targets_min": 1389 + }, + { + "epoch": 1.7257462686567164, + "grad_norm": 0.6602455517310905, + "learning_rate": 3.7455373756185276e-05, + "loss": 0.5176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24381765723228455, + "step": 925, + "valid_targets_mean": 3100.6, + "valid_targets_min": 1163 + }, + { + "epoch": 1.7350746268656716, + "grad_norm": 0.6479342508090896, + "learning_rate": 3.740976085757495e-05, + "loss": 0.508, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.290170282125473, + "step": 930, + "valid_targets_mean": 3588.8, + "valid_targets_min": 804 + }, + { + "epoch": 1.7444029850746268, + "grad_norm": 0.704484935116519, + "learning_rate": 3.736377105830259e-05, + "loss": 0.5459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3132808804512024, + "step": 935, + "valid_targets_mean": 3918.0, + "valid_targets_min": 1117 + }, + { + "epoch": 1.7537313432835822, + "grad_norm": 0.6733346721826577, + "learning_rate": 3.731740535399278e-05, + "loss": 0.5267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2909555435180664, + "step": 940, + "valid_targets_mean": 3299.9, + "valid_targets_min": 662 + }, + { + "epoch": 1.7630597014925373, + "grad_norm": 0.6710556888506656, + "learning_rate": 3.7270664748407985e-05, + "loss": 0.5194, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34600740671157837, + "step": 945, + "valid_targets_mean": 3741.2, + "valid_targets_min": 1129 + }, + { + "epoch": 1.7723880597014925, + "grad_norm": 0.6259514941534328, + "learning_rate": 3.722355025342686e-05, + "loss": 0.526, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.346868097782135, + "step": 950, + "valid_targets_mean": 4355.9, + "valid_targets_min": 1138 + }, + { + "epoch": 1.7817164179104479, + "grad_norm": 0.6863877629730316, + "learning_rate": 3.7176062889022296e-05, + "loss": 0.5453, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23799946904182434, + "step": 955, + "valid_targets_mean": 2967.8, + "valid_targets_min": 597 + }, + { + "epoch": 1.7910447761194028, + "grad_norm": 0.6419031733045666, + "learning_rate": 3.7128203683239384e-05, + "loss": 0.5628, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.39277732372283936, + "step": 960, + "valid_targets_mean": 4363.8, + "valid_targets_min": 903 + }, + { + "epoch": 1.8003731343283582, + "grad_norm": 0.692108057203365, + "learning_rate": 3.7079973672173136e-05, + "loss": 0.535, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32965022325515747, + "step": 965, + "valid_targets_mean": 3279.0, + "valid_targets_min": 1346 + }, + { + "epoch": 1.8097014925373134, + "grad_norm": 0.6286116196936884, + "learning_rate": 3.703137389994606e-05, + "loss": 0.4916, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22295673191547394, + "step": 970, + "valid_targets_mean": 3213.0, + "valid_targets_min": 1399 + }, + { + "epoch": 1.8190298507462686, + "grad_norm": 0.6525695959453189, + "learning_rate": 3.698240541868554e-05, + "loss": 0.5622, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.330750972032547, + "step": 975, + "valid_targets_mean": 3907.9, + "valid_targets_min": 1085 + }, + { + "epoch": 1.828358208955224, + "grad_norm": 0.7053578032830937, + "learning_rate": 3.693306928850109e-05, + "loss": 0.4906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25775429606437683, + "step": 980, + "valid_targets_mean": 2898.6, + "valid_targets_min": 1064 + }, + { + "epoch": 1.837686567164179, + "grad_norm": 0.6433038732128293, + "learning_rate": 3.688336657746137e-05, + "loss": 0.5296, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42391300201416016, + "step": 985, + "valid_targets_mean": 4564.8, + "valid_targets_min": 1470 + }, + { + "epoch": 1.8470149253731343, + "grad_norm": 0.6424701817752532, + "learning_rate": 3.683329836157111e-05, + "loss": 0.5396, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24525247514247894, + "step": 990, + "valid_targets_mean": 2879.2, + "valid_targets_min": 1077 + }, + { + "epoch": 1.8563432835820897, + "grad_norm": 0.6608009034763483, + "learning_rate": 3.678286572474776e-05, + "loss": 0.5106, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28463658690452576, + "step": 995, + "valid_targets_mean": 3474.4, + "valid_targets_min": 670 + }, + { + "epoch": 1.8656716417910446, + "grad_norm": 0.57720451658567, + "learning_rate": 3.673206975879804e-05, + "loss": 0.5101, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34058770537376404, + "step": 1000, + "valid_targets_mean": 4883.0, + "valid_targets_min": 1674 + }, + { + "epoch": 1.875, + "grad_norm": 0.6596801156484875, + "learning_rate": 3.668091156339435e-05, + "loss": 0.5369, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22180920839309692, + "step": 1005, + "valid_targets_mean": 2621.1, + "valid_targets_min": 594 + }, + { + "epoch": 1.8843283582089554, + "grad_norm": 0.5557907162669783, + "learning_rate": 3.662939224605091e-05, + "loss": 0.5012, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26842308044433594, + "step": 1010, + "valid_targets_mean": 5005.4, + "valid_targets_min": 1288 + }, + { + "epoch": 1.8936567164179103, + "grad_norm": 0.6335525248090774, + "learning_rate": 3.657751292209979e-05, + "loss": 0.5007, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24365246295928955, + "step": 1015, + "valid_targets_mean": 3732.0, + "valid_targets_min": 1214 + }, + { + "epoch": 1.9029850746268657, + "grad_norm": 0.5086468186696038, + "learning_rate": 3.65252747146668e-05, + "loss": 0.4933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2148883193731308, + "step": 1020, + "valid_targets_mean": 5148.0, + "valid_targets_min": 466 + }, + { + "epoch": 1.912313432835821, + "grad_norm": 0.6511561889056152, + "learning_rate": 3.647267875464714e-05, + "loss": 0.5149, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.283741295337677, + "step": 1025, + "valid_targets_mean": 3377.8, + "valid_targets_min": 1910 + }, + { + "epoch": 1.921641791044776, + "grad_norm": 0.566821738221381, + "learning_rate": 3.6419726180680914e-05, + "loss": 0.52, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22616514563560486, + "step": 1030, + "valid_targets_mean": 3953.8, + "valid_targets_min": 1879 + }, + { + "epoch": 1.9309701492537314, + "grad_norm": 0.5723425429689769, + "learning_rate": 3.636641813912851e-05, + "loss": 0.5329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19697515666484833, + "step": 1035, + "valid_targets_mean": 3088.0, + "valid_targets_min": 1851 + }, + { + "epoch": 1.9402985074626866, + "grad_norm": 0.6842233524447217, + "learning_rate": 3.631275578404578e-05, + "loss": 0.5328, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24774794280529022, + "step": 1040, + "valid_targets_mean": 3153.1, + "valid_targets_min": 2244 + }, + { + "epoch": 1.9496268656716418, + "grad_norm": 0.6075662099902452, + "learning_rate": 3.6258740277159014e-05, + "loss": 0.5352, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24449366331100464, + "step": 1045, + "valid_targets_mean": 3260.4, + "valid_targets_min": 1364 + }, + { + "epoch": 1.9589552238805972, + "grad_norm": 0.571434209357847, + "learning_rate": 3.6204372787839844e-05, + "loss": 0.5113, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22539398074150085, + "step": 1050, + "valid_targets_mean": 3225.4, + "valid_targets_min": 1551 + }, + { + "epoch": 1.9682835820895521, + "grad_norm": 0.7085509370310916, + "learning_rate": 3.614965449307989e-05, + "loss": 0.564, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3991805911064148, + "step": 1055, + "valid_targets_mean": 3845.9, + "valid_targets_min": 1744 + }, + { + "epoch": 1.9776119402985075, + "grad_norm": 0.6682551651142908, + "learning_rate": 3.609458657746531e-05, + "loss": 0.5365, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40009605884552, + "step": 1060, + "valid_targets_mean": 4338.5, + "valid_targets_min": 1944 + }, + { + "epoch": 1.9869402985074627, + "grad_norm": 0.6632898611224972, + "learning_rate": 3.603917023315111e-05, + "loss": 0.5223, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21475926041603088, + "step": 1065, + "valid_targets_mean": 2839.5, + "valid_targets_min": 402 + }, + { + "epoch": 1.9962686567164178, + "grad_norm": 0.5900696293527807, + "learning_rate": 3.598340665983539e-05, + "loss": 0.4822, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2842358648777008, + "step": 1070, + "valid_targets_mean": 4760.9, + "valid_targets_min": 768 + }, + { + "epoch": 2.0055970149253732, + "grad_norm": 0.6795321301741746, + "learning_rate": 3.592729706473333e-05, + "loss": 0.5024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2950167655944824, + "step": 1075, + "valid_targets_mean": 3725.5, + "valid_targets_min": 1344 + }, + { + "epoch": 2.014925373134328, + "grad_norm": 0.6700156452114036, + "learning_rate": 3.587084266255108e-05, + "loss": 0.5409, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31132829189300537, + "step": 1080, + "valid_targets_mean": 3652.6, + "valid_targets_min": 1406 + }, + { + "epoch": 2.0242537313432836, + "grad_norm": 0.5519808361597551, + "learning_rate": 3.581404467545946e-05, + "loss": 0.4673, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22447973489761353, + "step": 1085, + "valid_targets_mean": 4872.8, + "valid_targets_min": 961 + }, + { + "epoch": 2.033582089552239, + "grad_norm": 0.7133361779107861, + "learning_rate": 3.575690433306748e-05, + "loss": 0.4907, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2486574351787567, + "step": 1090, + "valid_targets_mean": 3758.1, + "valid_targets_min": 1161 + }, + { + "epoch": 2.042910447761194, + "grad_norm": 0.7138176459876374, + "learning_rate": 3.5699422872395735e-05, + "loss": 0.4737, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3353801965713501, + "step": 1095, + "valid_targets_mean": 4111.1, + "valid_targets_min": 725 + }, + { + "epoch": 2.0522388059701493, + "grad_norm": 0.5637860857369189, + "learning_rate": 3.564160153784965e-05, + "loss": 0.5171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28206759691238403, + "step": 1100, + "valid_targets_mean": 4734.6, + "valid_targets_min": 1752 + }, + { + "epoch": 2.0615671641791047, + "grad_norm": 0.6291197538037886, + "learning_rate": 3.5583441581192484e-05, + "loss": 0.4906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3299903869628906, + "step": 1105, + "valid_targets_mean": 4238.4, + "valid_targets_min": 2300 + }, + { + "epoch": 2.0708955223880596, + "grad_norm": 0.6194688488371134, + "learning_rate": 3.5524944261518295e-05, + "loss": 0.5119, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2566660940647125, + "step": 1110, + "valid_targets_mean": 4230.5, + "valid_targets_min": 1510 + }, + { + "epoch": 2.080223880597015, + "grad_norm": 0.6812045835658691, + "learning_rate": 3.546611084522462e-05, + "loss": 0.4977, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23873913288116455, + "step": 1115, + "valid_targets_mean": 3793.5, + "valid_targets_min": 1215 + }, + { + "epoch": 2.08955223880597, + "grad_norm": 0.671797040912562, + "learning_rate": 3.540694260598512e-05, + "loss": 0.4752, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19716548919677734, + "step": 1120, + "valid_targets_mean": 3814.0, + "valid_targets_min": 902 + }, + { + "epoch": 2.0988805970149254, + "grad_norm": 0.7364245828398875, + "learning_rate": 3.5347440824721946e-05, + "loss": 0.5074, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27472764253616333, + "step": 1125, + "valid_targets_mean": 3250.9, + "valid_targets_min": 1450 + }, + { + "epoch": 2.1082089552238807, + "grad_norm": 0.6253821166296961, + "learning_rate": 3.5287606789578046e-05, + "loss": 0.5059, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12621966004371643, + "step": 1130, + "valid_targets_mean": 2535.1, + "valid_targets_min": 1089 + }, + { + "epoch": 2.1175373134328357, + "grad_norm": 0.5836584723225509, + "learning_rate": 3.522744179588928e-05, + "loss": 0.491, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17364773154258728, + "step": 1135, + "valid_targets_mean": 3252.0, + "valid_targets_min": 857 + }, + { + "epoch": 2.126865671641791, + "grad_norm": 0.627176985631778, + "learning_rate": 3.516694714615637e-05, + "loss": 0.5014, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30840474367141724, + "step": 1140, + "valid_targets_mean": 5415.8, + "valid_targets_min": 1047 + }, + { + "epoch": 2.1361940298507465, + "grad_norm": 0.5736168934198965, + "learning_rate": 3.510612415001668e-05, + "loss": 0.4995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33349835872650146, + "step": 1145, + "valid_targets_mean": 5942.2, + "valid_targets_min": 1989 + }, + { + "epoch": 2.1455223880597014, + "grad_norm": 0.6771747319827419, + "learning_rate": 3.50449741242159e-05, + "loss": 0.4825, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2897084355354309, + "step": 1150, + "valid_targets_mean": 4179.0, + "valid_targets_min": 2159 + }, + { + "epoch": 2.154850746268657, + "grad_norm": 0.7042603235180996, + "learning_rate": 3.4983498392579526e-05, + "loss": 0.5049, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30181655287742615, + "step": 1155, + "valid_targets_mean": 3311.4, + "valid_targets_min": 1607 + }, + { + "epoch": 2.1641791044776117, + "grad_norm": 0.9016289058153439, + "learning_rate": 3.492169828598419e-05, + "loss": 0.5105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20479640364646912, + "step": 1160, + "valid_targets_mean": 2514.2, + "valid_targets_min": 888 + }, + { + "epoch": 2.173507462686567, + "grad_norm": 0.5837347162818661, + "learning_rate": 3.485957514232886e-05, + "loss": 0.527, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23902499675750732, + "step": 1165, + "valid_targets_mean": 3897.0, + "valid_targets_min": 1226 + }, + { + "epoch": 2.1828358208955225, + "grad_norm": 0.6674583634742567, + "learning_rate": 3.479713030650587e-05, + "loss": 0.529, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3221573829650879, + "step": 1170, + "valid_targets_mean": 4660.2, + "valid_targets_min": 2286 + }, + { + "epoch": 2.1921641791044775, + "grad_norm": 0.6101862059947767, + "learning_rate": 3.4734365130371816e-05, + "loss": 0.4656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22524692118167877, + "step": 1175, + "valid_targets_mean": 3418.2, + "valid_targets_min": 1242 + }, + { + "epoch": 2.201492537313433, + "grad_norm": 0.5803670171067227, + "learning_rate": 3.467128097271828e-05, + "loss": 0.4968, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20585867762565613, + "step": 1180, + "valid_targets_mean": 4331.2, + "valid_targets_min": 1596 + }, + { + "epoch": 2.2108208955223883, + "grad_norm": 0.6747048569976891, + "learning_rate": 3.46078791992424e-05, + "loss": 0.5111, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2578763961791992, + "step": 1185, + "valid_targets_mean": 3737.5, + "valid_targets_min": 1663 + }, + { + "epoch": 2.220149253731343, + "grad_norm": 0.590262455911102, + "learning_rate": 3.454416118251734e-05, + "loss": 0.4717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27388307452201843, + "step": 1190, + "valid_targets_mean": 4674.6, + "valid_targets_min": 1160 + }, + { + "epoch": 2.2294776119402986, + "grad_norm": 0.7094233697504353, + "learning_rate": 3.448012830196255e-05, + "loss": 0.498, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22383376955986023, + "step": 1195, + "valid_targets_mean": 2702.1, + "valid_targets_min": 1129 + }, + { + "epoch": 2.2388059701492535, + "grad_norm": 0.7050924641495339, + "learning_rate": 3.441578194381389e-05, + "loss": 0.4906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2757047414779663, + "step": 1200, + "valid_targets_mean": 3225.6, + "valid_targets_min": 1703 + }, + { + "epoch": 2.248134328358209, + "grad_norm": 0.5834766190678712, + "learning_rate": 3.435112350109367e-05, + "loss": 0.4856, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17494474351406097, + "step": 1205, + "valid_targets_mean": 3320.1, + "valid_targets_min": 1847 + }, + { + "epoch": 2.2574626865671643, + "grad_norm": 0.6221829969728038, + "learning_rate": 3.428615437358043e-05, + "loss": 0.4933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2225055992603302, + "step": 1210, + "valid_targets_mean": 3775.2, + "valid_targets_min": 639 + }, + { + "epoch": 2.2667910447761193, + "grad_norm": 0.7735507468748547, + "learning_rate": 3.422087596777869e-05, + "loss": 0.5027, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2231864631175995, + "step": 1215, + "valid_targets_mean": 2963.2, + "valid_targets_min": 1787 + }, + { + "epoch": 2.2761194029850746, + "grad_norm": 0.6125358674732613, + "learning_rate": 3.415528969688846e-05, + "loss": 0.4796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2681494355201721, + "step": 1220, + "valid_targets_mean": 4630.8, + "valid_targets_min": 402 + }, + { + "epoch": 2.28544776119403, + "grad_norm": 0.7300973067471473, + "learning_rate": 3.408939698077468e-05, + "loss": 0.4555, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23674388229846954, + "step": 1225, + "valid_targets_mean": 3540.2, + "valid_targets_min": 654 + }, + { + "epoch": 2.294776119402985, + "grad_norm": 0.6179957595566913, + "learning_rate": 3.402319924593645e-05, + "loss": 0.4917, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21130673587322235, + "step": 1230, + "valid_targets_mean": 4138.9, + "valid_targets_min": 774 + }, + { + "epoch": 2.3041044776119404, + "grad_norm": 0.6124235868995113, + "learning_rate": 3.395669792547618e-05, + "loss": 0.4669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2290029227733612, + "step": 1235, + "valid_targets_mean": 4125.1, + "valid_targets_min": 1378 + }, + { + "epoch": 2.3134328358208958, + "grad_norm": 0.7746413796509022, + "learning_rate": 3.388989445906853e-05, + "loss": 0.4901, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18875160813331604, + "step": 1240, + "valid_targets_mean": 2358.1, + "valid_targets_min": 452 + }, + { + "epoch": 2.3227611940298507, + "grad_norm": 0.6156818174509552, + "learning_rate": 3.382279029292928e-05, + "loss": 0.4864, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22659897804260254, + "step": 1245, + "valid_targets_mean": 3925.2, + "valid_targets_min": 375 + }, + { + "epoch": 2.332089552238806, + "grad_norm": 0.70256833923607, + "learning_rate": 3.3755386879783985e-05, + "loss": 0.4627, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2271793633699417, + "step": 1250, + "valid_targets_mean": 2551.8, + "valid_targets_min": 811 + }, + { + "epoch": 2.341417910447761, + "grad_norm": 0.707704048181143, + "learning_rate": 3.368768567883655e-05, + "loss": 0.4742, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2715616822242737, + "step": 1255, + "valid_targets_mean": 2697.1, + "valid_targets_min": 1609 + }, + { + "epoch": 2.3507462686567164, + "grad_norm": 0.6374055411905923, + "learning_rate": 3.3619688155737646e-05, + "loss": 0.4907, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22505158185958862, + "step": 1260, + "valid_targets_mean": 3017.2, + "valid_targets_min": 829 + }, + { + "epoch": 2.360074626865672, + "grad_norm": 0.6804013627450715, + "learning_rate": 3.355139578255294e-05, + "loss": 0.4711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29315218329429626, + "step": 1265, + "valid_targets_mean": 3606.0, + "valid_targets_min": 1308 + }, + { + "epoch": 2.3694029850746268, + "grad_norm": 0.5778617602796682, + "learning_rate": 3.348281003773127e-05, + "loss": 0.5177, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24891561269760132, + "step": 1270, + "valid_targets_mean": 3757.5, + "valid_targets_min": 476 + }, + { + "epoch": 2.378731343283582, + "grad_norm": 0.5700485688550925, + "learning_rate": 3.3413932406072626e-05, + "loss": 0.5027, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18104222416877747, + "step": 1275, + "valid_targets_mean": 3743.9, + "valid_targets_min": 1695 + }, + { + "epoch": 2.388059701492537, + "grad_norm": 0.6303961810823999, + "learning_rate": 3.3344764378695984e-05, + "loss": 0.5029, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2416057586669922, + "step": 1280, + "valid_targets_mean": 3173.9, + "valid_targets_min": 1648 + }, + { + "epoch": 2.3973880597014925, + "grad_norm": 0.6074070866782983, + "learning_rate": 3.3275307453007066e-05, + "loss": 0.4866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32136327028274536, + "step": 1285, + "valid_targets_mean": 5180.0, + "valid_targets_min": 501 + }, + { + "epoch": 2.406716417910448, + "grad_norm": 0.5174151663898331, + "learning_rate": 3.3205563132665876e-05, + "loss": 0.5012, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2086087465286255, + "step": 1290, + "valid_targets_mean": 4505.6, + "valid_targets_min": 1785 + }, + { + "epoch": 2.416044776119403, + "grad_norm": 0.6811709973294688, + "learning_rate": 3.313553292755419e-05, + "loss": 0.489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25120699405670166, + "step": 1295, + "valid_targets_mean": 3540.0, + "valid_targets_min": 1779 + }, + { + "epoch": 2.425373134328358, + "grad_norm": 0.6968352741136534, + "learning_rate": 3.306521835374284e-05, + "loss": 0.4889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23758551478385925, + "step": 1300, + "valid_targets_mean": 2756.1, + "valid_targets_min": 1442 + }, + { + "epoch": 2.4347014925373136, + "grad_norm": 0.620041563113699, + "learning_rate": 3.2994620933458896e-05, + "loss": 0.4847, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24452653527259827, + "step": 1305, + "valid_targets_mean": 3382.6, + "valid_targets_min": 369 + }, + { + "epoch": 2.4440298507462686, + "grad_norm": 0.6847751458425498, + "learning_rate": 3.292374219505274e-05, + "loss": 0.4819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2521475553512573, + "step": 1310, + "valid_targets_mean": 2660.2, + "valid_targets_min": 501 + }, + { + "epoch": 2.453358208955224, + "grad_norm": 0.6961134731916392, + "learning_rate": 3.285258367296496e-05, + "loss": 0.5311, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2897127866744995, + "step": 1315, + "valid_targets_mean": 6153.4, + "valid_targets_min": 1367 + }, + { + "epoch": 2.4626865671641793, + "grad_norm": 0.592079963749521, + "learning_rate": 3.278114690769307e-05, + "loss": 0.5187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32689380645751953, + "step": 1320, + "valid_targets_mean": 4144.6, + "valid_targets_min": 1334 + }, + { + "epoch": 2.4720149253731343, + "grad_norm": 0.6206416060392662, + "learning_rate": 3.2709433445758304e-05, + "loss": 0.4869, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2550956904888153, + "step": 1325, + "valid_targets_mean": 4343.4, + "valid_targets_min": 855 + }, + { + "epoch": 2.4813432835820897, + "grad_norm": 0.6456650989184988, + "learning_rate": 3.263744483967198e-05, + "loss": 0.4655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1758570820093155, + "step": 1330, + "valid_targets_mean": 2506.4, + "valid_targets_min": 1088 + }, + { + "epoch": 2.4906716417910446, + "grad_norm": 0.6871147518904734, + "learning_rate": 3.2565182647902015e-05, + "loss": 0.4912, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18231484293937683, + "step": 1335, + "valid_targets_mean": 2167.0, + "valid_targets_min": 855 + }, + { + "epoch": 2.5, + "grad_norm": 0.5508902183149501, + "learning_rate": 3.249264843483909e-05, + "loss": 0.5031, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1539400815963745, + "step": 1340, + "valid_targets_mean": 3702.0, + "valid_targets_min": 1257 + }, + { + "epoch": 2.5093283582089554, + "grad_norm": 0.711039764166237, + "learning_rate": 3.2419843770762836e-05, + "loss": 0.5416, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18896105885505676, + "step": 1345, + "valid_targets_mean": 2290.5, + "valid_targets_min": 1544 + }, + { + "epoch": 2.5186567164179103, + "grad_norm": 0.6457640292712167, + "learning_rate": 3.2346770231807843e-05, + "loss": 0.4774, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3772604465484619, + "step": 1350, + "valid_targets_mean": 4927.9, + "valid_targets_min": 1753 + }, + { + "epoch": 2.5279850746268657, + "grad_norm": 0.6804389911685375, + "learning_rate": 3.227342939992951e-05, + "loss": 0.5156, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2712757885456085, + "step": 1355, + "valid_targets_mean": 3482.5, + "valid_targets_min": 457 + }, + { + "epoch": 2.5373134328358207, + "grad_norm": 0.6131309330428244, + "learning_rate": 3.2199822862869803e-05, + "loss": 0.4677, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21193549036979675, + "step": 1360, + "valid_targets_mean": 3372.6, + "valid_targets_min": 1099 + }, + { + "epoch": 2.546641791044776, + "grad_norm": 0.7001543112102687, + "learning_rate": 3.212595221412291e-05, + "loss": 0.4918, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18551722168922424, + "step": 1365, + "valid_targets_mean": 2337.2, + "valid_targets_min": 893 + }, + { + "epoch": 2.5559701492537314, + "grad_norm": 0.5873958410222707, + "learning_rate": 3.2051819052900706e-05, + "loss": 0.4996, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2074936032295227, + "step": 1370, + "valid_targets_mean": 3923.1, + "valid_targets_min": 1609 + }, + { + "epoch": 2.5652985074626864, + "grad_norm": 0.8078702496254674, + "learning_rate": 3.1977424984098154e-05, + "loss": 0.4767, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23863083124160767, + "step": 1375, + "valid_targets_mean": 3540.8, + "valid_targets_min": 894 + }, + { + "epoch": 2.574626865671642, + "grad_norm": 0.7164836881539058, + "learning_rate": 3.190277161825857e-05, + "loss": 0.501, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2595667541027069, + "step": 1380, + "valid_targets_mean": 3372.1, + "valid_targets_min": 1798 + }, + { + "epoch": 2.583955223880597, + "grad_norm": 0.6511280392576557, + "learning_rate": 3.182786057153871e-05, + "loss": 0.5063, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.325607568025589, + "step": 1385, + "valid_targets_mean": 3256.8, + "valid_targets_min": 1063 + }, + { + "epoch": 2.593283582089552, + "grad_norm": 0.6480325819326036, + "learning_rate": 3.175269346567386e-05, + "loss": 0.4559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20964059233665466, + "step": 1390, + "valid_targets_mean": 2873.8, + "valid_targets_min": 1025 + }, + { + "epoch": 2.6026119402985075, + "grad_norm": 0.6739136932028859, + "learning_rate": 3.167727192794265e-05, + "loss": 0.5141, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2113959640264511, + "step": 1395, + "valid_targets_mean": 3000.4, + "valid_targets_min": 1270 + }, + { + "epoch": 2.611940298507463, + "grad_norm": 0.6099397018654773, + "learning_rate": 3.1601597591131864e-05, + "loss": 0.4684, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22376658022403717, + "step": 1400, + "valid_targets_mean": 3665.1, + "valid_targets_min": 724 + }, + { + "epoch": 2.621268656716418, + "grad_norm": 0.6580728466619485, + "learning_rate": 3.152567209350109e-05, + "loss": 0.4915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26452547311782837, + "step": 1405, + "valid_targets_mean": 3563.2, + "valid_targets_min": 1089 + }, + { + "epoch": 2.6305970149253732, + "grad_norm": 0.7663779704239244, + "learning_rate": 3.144949707874726e-05, + "loss": 0.4866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3006416857242584, + "step": 1410, + "valid_targets_mean": 2725.8, + "valid_targets_min": 726 + }, + { + "epoch": 2.6399253731343286, + "grad_norm": 0.6601676798664754, + "learning_rate": 3.137307419596904e-05, + "loss": 0.4998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25848090648651123, + "step": 1415, + "valid_targets_mean": 2898.8, + "valid_targets_min": 1245 + }, + { + "epoch": 2.6492537313432836, + "grad_norm": 0.5414547000360638, + "learning_rate": 3.129640509963114e-05, + "loss": 0.5035, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16674742102622986, + "step": 1420, + "valid_targets_mean": 4123.4, + "valid_targets_min": 1264 + }, + { + "epoch": 2.658582089552239, + "grad_norm": 0.701858019472102, + "learning_rate": 3.121949144952853e-05, + "loss": 0.4857, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2097078561782837, + "step": 1425, + "valid_targets_mean": 2161.5, + "valid_targets_min": 743 + }, + { + "epoch": 2.667910447761194, + "grad_norm": 0.6009021035693228, + "learning_rate": 3.1142334910750426e-05, + "loss": 0.5153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20608268678188324, + "step": 1430, + "valid_targets_mean": 3786.6, + "valid_targets_min": 325 + }, + { + "epoch": 2.6772388059701493, + "grad_norm": 0.6632875521320487, + "learning_rate": 3.1064937153644366e-05, + "loss": 0.503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23304741084575653, + "step": 1435, + "valid_targets_mean": 2971.9, + "valid_targets_min": 1311 + }, + { + "epoch": 2.6865671641791042, + "grad_norm": 0.7285800630216194, + "learning_rate": 3.098729985377992e-05, + "loss": 0.5105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34828880429267883, + "step": 1440, + "valid_targets_mean": 3886.1, + "valid_targets_min": 903 + }, + { + "epoch": 2.6958955223880596, + "grad_norm": 0.6718117361757154, + "learning_rate": 3.090942469191251e-05, + "loss": 0.4605, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20975592732429504, + "step": 1445, + "valid_targets_mean": 2620.1, + "valid_targets_min": 1032 + }, + { + "epoch": 2.705223880597015, + "grad_norm": 0.6602204683123232, + "learning_rate": 3.083131335394696e-05, + "loss": 0.493, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23395368456840515, + "step": 1450, + "valid_targets_mean": 3166.9, + "valid_targets_min": 1196 + }, + { + "epoch": 2.71455223880597, + "grad_norm": 1.0138656170234317, + "learning_rate": 3.075296753090105e-05, + "loss": 0.5031, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28172895312309265, + "step": 1455, + "valid_targets_mean": 5620.6, + "valid_targets_min": 1485 + }, + { + "epoch": 2.7238805970149254, + "grad_norm": 0.6090420112266862, + "learning_rate": 3.0674388918868876e-05, + "loss": 0.4703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2475488930940628, + "step": 1460, + "valid_targets_mean": 3834.6, + "valid_targets_min": 808 + }, + { + "epoch": 2.7332089552238807, + "grad_norm": 0.7955105676221782, + "learning_rate": 3.0595579218984124e-05, + "loss": 0.5103, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20074407756328583, + "step": 1465, + "valid_targets_mean": 3846.5, + "valid_targets_min": 1199 + }, + { + "epoch": 2.7425373134328357, + "grad_norm": 0.6584297635453287, + "learning_rate": 3.0516540137383287e-05, + "loss": 0.4848, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3010520339012146, + "step": 1470, + "valid_targets_mean": 3945.9, + "valid_targets_min": 1266 + }, + { + "epoch": 2.751865671641791, + "grad_norm": 0.6537176275220778, + "learning_rate": 3.0437273385168677e-05, + "loss": 0.4981, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3279004693031311, + "step": 1475, + "valid_targets_mean": 3718.4, + "valid_targets_min": 1182 + }, + { + "epoch": 2.7611940298507465, + "grad_norm": 0.6962141019375813, + "learning_rate": 3.0357780678371413e-05, + "loss": 0.4851, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2941030263900757, + "step": 1480, + "valid_targets_mean": 3283.8, + "valid_targets_min": 1519 + }, + { + "epoch": 2.7705223880597014, + "grad_norm": 0.6620586132345421, + "learning_rate": 3.027806373791426e-05, + "loss": 0.4653, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30104300379753113, + "step": 1485, + "valid_targets_mean": 3244.5, + "valid_targets_min": 435 + }, + { + "epoch": 2.779850746268657, + "grad_norm": 0.6191132428377458, + "learning_rate": 3.0198124289574378e-05, + "loss": 0.4661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31103208661079407, + "step": 1490, + "valid_targets_mean": 4477.1, + "valid_targets_min": 1665 + }, + { + "epoch": 2.789179104477612, + "grad_norm": 0.7294073633888214, + "learning_rate": 3.011796406394596e-05, + "loss": 0.5151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21798336505889893, + "step": 1495, + "valid_targets_mean": 2673.2, + "valid_targets_min": 1070 + }, + { + "epoch": 2.798507462686567, + "grad_norm": 0.6153766824967245, + "learning_rate": 3.0037584796402767e-05, + "loss": 0.4883, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21332812309265137, + "step": 1500, + "valid_targets_mean": 3697.9, + "valid_targets_min": 1377 + }, + { + "epoch": 2.8078358208955225, + "grad_norm": 0.7184753622696983, + "learning_rate": 2.9956988227060536e-05, + "loss": 0.5217, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3174893856048584, + "step": 1505, + "valid_targets_mean": 2870.5, + "valid_targets_min": 1069 + }, + { + "epoch": 2.8171641791044775, + "grad_norm": 0.586297965696393, + "learning_rate": 2.9876176100739368e-05, + "loss": 0.5065, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26149946451187134, + "step": 1510, + "valid_targets_mean": 3866.9, + "valid_targets_min": 723 + }, + { + "epoch": 2.826492537313433, + "grad_norm": 0.6378897543495345, + "learning_rate": 2.979515016692589e-05, + "loss": 0.4586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2185828685760498, + "step": 1515, + "valid_targets_mean": 3313.0, + "valid_targets_min": 961 + }, + { + "epoch": 2.835820895522388, + "grad_norm": 0.6635766086993317, + "learning_rate": 2.9713912179735404e-05, + "loss": 0.4819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26075634360313416, + "step": 1520, + "valid_targets_mean": 3274.4, + "valid_targets_min": 1291 + }, + { + "epoch": 2.845149253731343, + "grad_norm": 0.6324304895234182, + "learning_rate": 2.9632463897873935e-05, + "loss": 0.4723, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15830416977405548, + "step": 1525, + "valid_targets_mean": 2200.9, + "valid_targets_min": 756 + }, + { + "epoch": 2.8544776119402986, + "grad_norm": 0.6577101175997724, + "learning_rate": 2.9550807084600134e-05, + "loss": 0.5032, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2642113268375397, + "step": 1530, + "valid_targets_mean": 3048.1, + "valid_targets_min": 1288 + }, + { + "epoch": 2.8638059701492535, + "grad_norm": 0.5739447580500147, + "learning_rate": 2.946894350768709e-05, + "loss": 0.4886, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28352904319763184, + "step": 1535, + "valid_targets_mean": 4393.1, + "valid_targets_min": 1981 + }, + { + "epoch": 2.873134328358209, + "grad_norm": 0.6303845143845707, + "learning_rate": 2.9386874939384103e-05, + "loss": 0.49, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23921620845794678, + "step": 1540, + "valid_targets_mean": 3351.8, + "valid_targets_min": 1687 + }, + { + "epoch": 2.8824626865671643, + "grad_norm": 0.5969723944635307, + "learning_rate": 2.9304603156378283e-05, + "loss": 0.4836, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3032902479171753, + "step": 1545, + "valid_targets_mean": 4229.8, + "valid_targets_min": 1891 + }, + { + "epoch": 2.8917910447761193, + "grad_norm": 0.5614183223258822, + "learning_rate": 2.9222129939756104e-05, + "loss": 0.5168, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28644824028015137, + "step": 1550, + "valid_targets_mean": 5066.0, + "valid_targets_min": 1806 + }, + { + "epoch": 2.9011194029850746, + "grad_norm": 0.650344048770152, + "learning_rate": 2.913945707496482e-05, + "loss": 0.4895, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27156633138656616, + "step": 1555, + "valid_targets_mean": 3273.4, + "valid_targets_min": 1504 + }, + { + "epoch": 2.91044776119403, + "grad_norm": 0.685036229365782, + "learning_rate": 2.9056586351773854e-05, + "loss": 0.471, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24835284054279327, + "step": 1560, + "valid_targets_mean": 3268.9, + "valid_targets_min": 865 + }, + { + "epoch": 2.919776119402985, + "grad_norm": 0.5800672139573582, + "learning_rate": 2.8973519564236024e-05, + "loss": 0.4799, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3277813196182251, + "step": 1565, + "valid_targets_mean": 5389.5, + "valid_targets_min": 1746 + }, + { + "epoch": 2.9291044776119404, + "grad_norm": 0.6974994480774259, + "learning_rate": 2.8890258510648685e-05, + "loss": 0.4801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20311880111694336, + "step": 1570, + "valid_targets_mean": 2409.4, + "valid_targets_min": 825 + }, + { + "epoch": 2.9384328358208958, + "grad_norm": 0.6113636930011851, + "learning_rate": 2.8806804993514852e-05, + "loss": 0.4695, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20716959238052368, + "step": 1575, + "valid_targets_mean": 3573.8, + "valid_targets_min": 2004 + }, + { + "epoch": 2.9477611940298507, + "grad_norm": 0.5947126470205408, + "learning_rate": 2.8723160819504133e-05, + "loss": 0.478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2597846984863281, + "step": 1580, + "valid_targets_mean": 3931.2, + "valid_targets_min": 1369 + }, + { + "epoch": 2.957089552238806, + "grad_norm": 0.5826649649084596, + "learning_rate": 2.8639327799413635e-05, + "loss": 0.4783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2456882745027542, + "step": 1585, + "valid_targets_mean": 4975.0, + "valid_targets_min": 749 + }, + { + "epoch": 2.966417910447761, + "grad_norm": 0.4977682763067925, + "learning_rate": 2.8555307748128752e-05, + "loss": 0.4676, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1954365074634552, + "step": 1590, + "valid_targets_mean": 3682.9, + "valid_targets_min": 1042 + }, + { + "epoch": 2.9757462686567164, + "grad_norm": 0.6364435674804109, + "learning_rate": 2.847110248458389e-05, + "loss": 0.4881, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2247084528207779, + "step": 1595, + "valid_targets_mean": 3431.2, + "valid_targets_min": 1435 + }, + { + "epoch": 2.9850746268656714, + "grad_norm": 0.5597563062640879, + "learning_rate": 2.838671383172307e-05, + "loss": 0.4796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2451513111591339, + "step": 1600, + "valid_targets_mean": 4081.1, + "valid_targets_min": 1139 + }, + { + "epoch": 2.9944029850746268, + "grad_norm": 0.5553277817672154, + "learning_rate": 2.830214361646049e-05, + "loss": 0.4917, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2299538254737854, + "step": 1605, + "valid_targets_mean": 3452.5, + "valid_targets_min": 948 + }, + { + "epoch": 3.003731343283582, + "grad_norm": 0.6769397410984466, + "learning_rate": 2.8217393669640944e-05, + "loss": 0.4777, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18012799322605133, + "step": 1610, + "valid_targets_mean": 2806.8, + "valid_targets_min": 1223 + }, + { + "epoch": 3.013059701492537, + "grad_norm": 0.6258966755810921, + "learning_rate": 2.8132465826000207e-05, + "loss": 0.4325, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21304531395435333, + "step": 1615, + "valid_targets_mean": 3987.6, + "valid_targets_min": 1931 + }, + { + "epoch": 3.0223880597014925, + "grad_norm": 1.4095443450495335, + "learning_rate": 2.8047361924125307e-05, + "loss": 0.456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1949651539325714, + "step": 1620, + "valid_targets_mean": 3917.6, + "valid_targets_min": 1745 + }, + { + "epoch": 3.031716417910448, + "grad_norm": 0.6984628961117604, + "learning_rate": 2.7962083806414727e-05, + "loss": 0.4691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3038066625595093, + "step": 1625, + "valid_targets_mean": 3949.0, + "valid_targets_min": 1819 + }, + { + "epoch": 3.041044776119403, + "grad_norm": 0.6447912070400428, + "learning_rate": 2.7876633319038514e-05, + "loss": 0.4593, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.289588987827301, + "step": 1630, + "valid_targets_mean": 4024.4, + "valid_targets_min": 1865 + }, + { + "epoch": 3.050373134328358, + "grad_norm": 0.6966825700390333, + "learning_rate": 2.779101231189831e-05, + "loss": 0.4487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21856652200222015, + "step": 1635, + "valid_targets_mean": 3307.5, + "valid_targets_min": 930 + }, + { + "epoch": 3.0597014925373136, + "grad_norm": 0.6277943045207356, + "learning_rate": 2.7705222638587308e-05, + "loss": 0.4865, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3393063545227051, + "step": 1640, + "valid_targets_mean": 4864.1, + "valid_targets_min": 1324 + }, + { + "epoch": 3.0690298507462686, + "grad_norm": 0.5459816910279144, + "learning_rate": 2.761926615635014e-05, + "loss": 0.4582, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22999955713748932, + "step": 1645, + "valid_targets_mean": 5024.9, + "valid_targets_min": 932 + }, + { + "epoch": 3.078358208955224, + "grad_norm": 0.6685922685904541, + "learning_rate": 2.7533144726042632e-05, + "loss": 0.4923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22103899717330933, + "step": 1650, + "valid_targets_mean": 3362.5, + "valid_targets_min": 1389 + }, + { + "epoch": 3.0876865671641793, + "grad_norm": 0.6617214080620449, + "learning_rate": 2.7446860212091548e-05, + "loss": 0.4886, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3098272383213043, + "step": 1655, + "valid_targets_mean": 4108.0, + "valid_targets_min": 1895 + }, + { + "epoch": 3.0970149253731343, + "grad_norm": 0.6673269579732327, + "learning_rate": 2.736041448245422e-05, + "loss": 0.4647, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20112240314483643, + "step": 1660, + "valid_targets_mean": 3023.4, + "valid_targets_min": 1158 + }, + { + "epoch": 3.1063432835820897, + "grad_norm": 0.6648024181550537, + "learning_rate": 2.7273809408578107e-05, + "loss": 0.4409, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2210225909948349, + "step": 1665, + "valid_targets_mean": 3647.8, + "valid_targets_min": 1565 + }, + { + "epoch": 3.1156716417910446, + "grad_norm": 0.5939315266621085, + "learning_rate": 2.7187046865360285e-05, + "loss": 0.4832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24101471900939941, + "step": 1670, + "valid_targets_mean": 4587.2, + "valid_targets_min": 2409 + }, + { + "epoch": 3.125, + "grad_norm": 0.7948939450733529, + "learning_rate": 2.7100128731106847e-05, + "loss": 0.4952, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29429101943969727, + "step": 1675, + "valid_targets_mean": 2791.1, + "valid_targets_min": 1484 + }, + { + "epoch": 3.1343283582089554, + "grad_norm": 0.8371224180169287, + "learning_rate": 2.7013056887492255e-05, + "loss": 0.4623, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1886288821697235, + "step": 1680, + "valid_targets_mean": 3679.9, + "valid_targets_min": 1295 + }, + { + "epoch": 3.1436567164179103, + "grad_norm": 0.658365599194327, + "learning_rate": 2.6925833219518596e-05, + "loss": 0.4574, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20357602834701538, + "step": 1685, + "valid_targets_mean": 3298.5, + "valid_targets_min": 1113 + }, + { + "epoch": 3.1529850746268657, + "grad_norm": 0.7724632536337025, + "learning_rate": 2.6838459615474758e-05, + "loss": 0.4685, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2704656720161438, + "step": 1690, + "valid_targets_mean": 2908.4, + "valid_targets_min": 1553 + }, + { + "epoch": 3.1623134328358207, + "grad_norm": 0.7779120141439028, + "learning_rate": 2.6750937966895603e-05, + "loss": 0.4484, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23353441059589386, + "step": 1695, + "valid_targets_mean": 2603.1, + "valid_targets_min": 1252 + }, + { + "epoch": 3.171641791044776, + "grad_norm": 0.6347415592880836, + "learning_rate": 2.666327016852094e-05, + "loss": 0.4592, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14353036880493164, + "step": 1700, + "valid_targets_mean": 2147.9, + "valid_targets_min": 1554 + }, + { + "epoch": 3.1809701492537314, + "grad_norm": 0.6732161040737858, + "learning_rate": 2.6575458118254576e-05, + "loss": 0.471, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16501805186271667, + "step": 1705, + "valid_targets_mean": 4134.8, + "valid_targets_min": 1052 + }, + { + "epoch": 3.1902985074626864, + "grad_norm": 0.6870699314919898, + "learning_rate": 2.648750371712319e-05, + "loss": 0.4889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2769854664802551, + "step": 1710, + "valid_targets_mean": 3483.0, + "valid_targets_min": 802 + }, + { + "epoch": 3.199626865671642, + "grad_norm": 0.6586654030946709, + "learning_rate": 2.6399408869235187e-05, + "loss": 0.4492, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21777082979679108, + "step": 1715, + "valid_targets_mean": 3572.0, + "valid_targets_min": 1480 + }, + { + "epoch": 3.208955223880597, + "grad_norm": 0.7423652292119057, + "learning_rate": 2.6311175481739473e-05, + "loss": 0.4636, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23949582874774933, + "step": 1720, + "valid_targets_mean": 3096.9, + "valid_targets_min": 1338 + }, + { + "epoch": 3.218283582089552, + "grad_norm": 0.5681028566856955, + "learning_rate": 2.6222805464784196e-05, + "loss": 0.4925, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21227262914180756, + "step": 1725, + "valid_targets_mean": 4152.2, + "valid_targets_min": 1744 + }, + { + "epoch": 3.2276119402985075, + "grad_norm": 0.6445644716218911, + "learning_rate": 2.6134300731475337e-05, + "loss": 0.4569, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3471381664276123, + "step": 1730, + "valid_targets_mean": 6039.4, + "valid_targets_min": 1858 + }, + { + "epoch": 3.236940298507463, + "grad_norm": 0.6232395841682372, + "learning_rate": 2.6045663197835352e-05, + "loss": 0.4482, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2452152967453003, + "step": 1735, + "valid_targets_mean": 3984.8, + "valid_targets_min": 1315 + }, + { + "epoch": 3.246268656716418, + "grad_norm": 0.6689013420316136, + "learning_rate": 2.5956894782761638e-05, + "loss": 0.4679, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20851817727088928, + "step": 1740, + "valid_targets_mean": 4113.1, + "valid_targets_min": 491 + }, + { + "epoch": 3.2555970149253732, + "grad_norm": 0.7817968039508213, + "learning_rate": 2.5867997407985063e-05, + "loss": 0.4582, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17310580611228943, + "step": 1745, + "valid_targets_mean": 2387.2, + "valid_targets_min": 277 + }, + { + "epoch": 3.264925373134328, + "grad_norm": 0.602786481494261, + "learning_rate": 2.577897299802828e-05, + "loss": 0.4797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2347017526626587, + "step": 1750, + "valid_targets_mean": 3963.0, + "valid_targets_min": 1524 + }, + { + "epoch": 3.2742537313432836, + "grad_norm": 0.6948755418045652, + "learning_rate": 2.568982348016412e-05, + "loss": 0.4729, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2813548147678375, + "step": 1755, + "valid_targets_mean": 3529.0, + "valid_targets_min": 667 + }, + { + "epoch": 3.283582089552239, + "grad_norm": 0.6508096199706745, + "learning_rate": 2.560055078437385e-05, + "loss": 0.4748, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15489208698272705, + "step": 1760, + "valid_targets_mean": 2464.4, + "valid_targets_min": 1436 + }, + { + "epoch": 3.292910447761194, + "grad_norm": 0.6394399661275905, + "learning_rate": 2.5511156843305393e-05, + "loss": 0.4742, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18780454993247986, + "step": 1765, + "valid_targets_mean": 2874.6, + "valid_targets_min": 570 + }, + { + "epoch": 3.3022388059701493, + "grad_norm": 0.5426952708824583, + "learning_rate": 2.542164359223148e-05, + "loss": 0.4334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20754581689834595, + "step": 1770, + "valid_targets_mean": 4450.2, + "valid_targets_min": 1586 + }, + { + "epoch": 3.3115671641791042, + "grad_norm": 0.7615058513597199, + "learning_rate": 2.5332012969007786e-05, + "loss": 0.46, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24364234507083893, + "step": 1775, + "valid_targets_mean": 2718.2, + "valid_targets_min": 866 + }, + { + "epoch": 3.3208955223880596, + "grad_norm": 0.6278435536211145, + "learning_rate": 2.5242266914030916e-05, + "loss": 0.4698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17533516883850098, + "step": 1780, + "valid_targets_mean": 2818.2, + "valid_targets_min": 1412 + }, + { + "epoch": 3.330223880597015, + "grad_norm": 0.6891913626690295, + "learning_rate": 2.5152407370196467e-05, + "loss": 0.4746, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2931355834007263, + "step": 1785, + "valid_targets_mean": 3828.2, + "valid_targets_min": 1255 + }, + { + "epoch": 3.33955223880597, + "grad_norm": 0.5512692030970632, + "learning_rate": 2.5062436282856926e-05, + "loss": 0.4137, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20585976541042328, + "step": 1790, + "valid_targets_mean": 4989.5, + "valid_targets_min": 918 + }, + { + "epoch": 3.3488805970149254, + "grad_norm": 0.6241846165679275, + "learning_rate": 2.4972355599779566e-05, + "loss": 0.4259, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24647779762744904, + "step": 1795, + "valid_targets_mean": 3453.9, + "valid_targets_min": 1088 + }, + { + "epoch": 3.3582089552238807, + "grad_norm": 0.6453164860447758, + "learning_rate": 2.4882167271104272e-05, + "loss": 0.4675, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2697363495826721, + "step": 1800, + "valid_targets_mean": 4913.6, + "valid_targets_min": 1627 + }, + { + "epoch": 3.3675373134328357, + "grad_norm": 0.7212735185697192, + "learning_rate": 2.4791873249301327e-05, + "loss": 0.481, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.157449409365654, + "step": 1805, + "valid_targets_mean": 2671.4, + "valid_targets_min": 477 + }, + { + "epoch": 3.376865671641791, + "grad_norm": 0.7818601591993207, + "learning_rate": 2.4701475489129152e-05, + "loss": 0.4598, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2335238754749298, + "step": 1810, + "valid_targets_mean": 3653.9, + "valid_targets_min": 1356 + }, + { + "epoch": 3.3861940298507465, + "grad_norm": 0.5599507320347956, + "learning_rate": 2.4610975947591988e-05, + "loss": 0.4507, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24904865026474, + "step": 1815, + "valid_targets_mean": 4975.5, + "valid_targets_min": 1951 + }, + { + "epoch": 3.3955223880597014, + "grad_norm": 0.6463106135557956, + "learning_rate": 2.452037658389749e-05, + "loss": 0.4549, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16001123189926147, + "step": 1820, + "valid_targets_mean": 2904.5, + "valid_targets_min": 835 + }, + { + "epoch": 3.404850746268657, + "grad_norm": 0.6738415085936987, + "learning_rate": 2.4429679359414386e-05, + "loss": 0.4769, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16198870539665222, + "step": 1825, + "valid_targets_mean": 2555.5, + "valid_targets_min": 602 + }, + { + "epoch": 3.4141791044776117, + "grad_norm": 0.6341956885226937, + "learning_rate": 2.433888623762994e-05, + "loss": 0.4503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1892627775669098, + "step": 1830, + "valid_targets_mean": 3953.6, + "valid_targets_min": 867 + }, + { + "epoch": 3.423507462686567, + "grad_norm": 0.5390211593848352, + "learning_rate": 2.42479991841075e-05, + "loss": 0.4556, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25461092591285706, + "step": 1835, + "valid_targets_mean": 5975.1, + "valid_targets_min": 2068 + }, + { + "epoch": 3.4328358208955225, + "grad_norm": 0.6896587441881304, + "learning_rate": 2.4157020166443913e-05, + "loss": 0.4623, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1682130992412567, + "step": 1840, + "valid_targets_mean": 2311.6, + "valid_targets_min": 881 + }, + { + "epoch": 3.4421641791044775, + "grad_norm": 0.5557591944781534, + "learning_rate": 2.4065951154226948e-05, + "loss": 0.4702, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24712495505809784, + "step": 1845, + "valid_targets_mean": 5027.4, + "valid_targets_min": 1809 + }, + { + "epoch": 3.451492537313433, + "grad_norm": 0.5482774907552572, + "learning_rate": 2.3974794118992652e-05, + "loss": 0.4714, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31354960799217224, + "step": 1850, + "valid_targets_mean": 6311.1, + "valid_targets_min": 1477 + }, + { + "epoch": 3.4608208955223883, + "grad_norm": 0.6798013414192975, + "learning_rate": 2.3883551034182647e-05, + "loss": 0.4685, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18115438520908356, + "step": 1855, + "valid_targets_mean": 2968.9, + "valid_targets_min": 796 + }, + { + "epoch": 3.470149253731343, + "grad_norm": 0.6739969102663479, + "learning_rate": 2.3792223875101446e-05, + "loss": 0.4578, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24926114082336426, + "step": 1860, + "valid_targets_mean": 3889.9, + "valid_targets_min": 1604 + }, + { + "epoch": 3.4794776119402986, + "grad_norm": 0.6493307135105036, + "learning_rate": 2.3700814618873662e-05, + "loss": 0.4533, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2681945562362671, + "step": 1865, + "valid_targets_mean": 4520.9, + "valid_targets_min": 2444 + }, + { + "epoch": 3.4888059701492535, + "grad_norm": 0.7397135105750157, + "learning_rate": 2.3609325244401212e-05, + "loss": 0.4763, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16068419814109802, + "step": 1870, + "valid_targets_mean": 2048.9, + "valid_targets_min": 786 + }, + { + "epoch": 3.498134328358209, + "grad_norm": 0.5449008803196878, + "learning_rate": 2.3517757732320484e-05, + "loss": 0.4462, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3325706720352173, + "step": 1875, + "valid_targets_mean": 6568.8, + "valid_targets_min": 1058 + }, + { + "epoch": 3.5074626865671643, + "grad_norm": 0.5880348418083413, + "learning_rate": 2.342611406495946e-05, + "loss": 0.4811, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20652616024017334, + "step": 1880, + "valid_targets_mean": 4391.2, + "valid_targets_min": 1013 + }, + { + "epoch": 3.5167910447761193, + "grad_norm": 0.6082289468736123, + "learning_rate": 2.333439622629476e-05, + "loss": 0.453, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23787692189216614, + "step": 1885, + "valid_targets_mean": 4300.1, + "valid_targets_min": 1325 + }, + { + "epoch": 3.5261194029850746, + "grad_norm": 0.5929306798569756, + "learning_rate": 2.3242606201908758e-05, + "loss": 0.4737, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20460283756256104, + "step": 1890, + "valid_targets_mean": 3846.1, + "valid_targets_min": 948 + }, + { + "epoch": 3.53544776119403, + "grad_norm": 0.6003797791367875, + "learning_rate": 2.3150745978946563e-05, + "loss": 0.4519, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19913290441036224, + "step": 1895, + "valid_targets_mean": 4068.2, + "valid_targets_min": 1301 + }, + { + "epoch": 3.544776119402985, + "grad_norm": 0.7027364822717801, + "learning_rate": 2.3058817546072976e-05, + "loss": 0.4539, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22909030318260193, + "step": 1900, + "valid_targets_mean": 2868.6, + "valid_targets_min": 961 + }, + { + "epoch": 3.5541044776119404, + "grad_norm": 0.5717990920557517, + "learning_rate": 2.2966822893429487e-05, + "loss": 0.4681, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2402801513671875, + "step": 1905, + "valid_targets_mean": 5167.6, + "valid_targets_min": 932 + }, + { + "epoch": 3.5634328358208958, + "grad_norm": 0.6711839169484709, + "learning_rate": 2.287476401259114e-05, + "loss": 0.4667, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19033388793468475, + "step": 1910, + "valid_targets_mean": 2537.2, + "valid_targets_min": 920 + }, + { + "epoch": 3.5727611940298507, + "grad_norm": 0.6254478599520235, + "learning_rate": 2.278264289652347e-05, + "loss": 0.4426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2854100465774536, + "step": 1915, + "valid_targets_mean": 3527.5, + "valid_targets_min": 1265 + }, + { + "epoch": 3.582089552238806, + "grad_norm": 0.6372796411736636, + "learning_rate": 2.269046153953931e-05, + "loss": 0.4701, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24537506699562073, + "step": 1920, + "valid_targets_mean": 4136.6, + "valid_targets_min": 1917 + }, + { + "epoch": 3.591417910447761, + "grad_norm": 0.6653402199306333, + "learning_rate": 2.2598221937255645e-05, + "loss": 0.478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26774513721466064, + "step": 1925, + "valid_targets_mean": 3324.2, + "valid_targets_min": 1567 + }, + { + "epoch": 3.6007462686567164, + "grad_norm": 0.7026134071207524, + "learning_rate": 2.2505926086550404e-05, + "loss": 0.4608, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.217864990234375, + "step": 1930, + "valid_targets_mean": 3260.9, + "valid_targets_min": 855 + }, + { + "epoch": 3.6100746268656714, + "grad_norm": 0.7143538856618112, + "learning_rate": 2.2413575985519226e-05, + "loss": 0.4507, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20595791935920715, + "step": 1935, + "valid_targets_mean": 2612.5, + "valid_targets_min": 686 + }, + { + "epoch": 3.6194029850746268, + "grad_norm": 0.6472256255263096, + "learning_rate": 2.2321173633432202e-05, + "loss": 0.4686, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20282460749149323, + "step": 1940, + "valid_targets_mean": 3795.2, + "valid_targets_min": 1646 + }, + { + "epoch": 3.628731343283582, + "grad_norm": 0.7154286581397117, + "learning_rate": 2.2228721030690603e-05, + "loss": 0.4314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.285696417093277, + "step": 1945, + "valid_targets_mean": 4864.6, + "valid_targets_min": 2221 + }, + { + "epoch": 3.638059701492537, + "grad_norm": 0.6484892359813883, + "learning_rate": 2.2136220178783557e-05, + "loss": 0.5067, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30165889859199524, + "step": 1950, + "valid_targets_mean": 4857.2, + "valid_targets_min": 2428 + }, + { + "epoch": 3.6473880597014925, + "grad_norm": 0.5786628370899648, + "learning_rate": 2.2043673080244744e-05, + "loss": 0.4273, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15055450797080994, + "step": 1955, + "valid_targets_mean": 3154.8, + "valid_targets_min": 1451 + }, + { + "epoch": 3.656716417910448, + "grad_norm": 0.6823982075710977, + "learning_rate": 2.1951081738609012e-05, + "loss": 0.483, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2150236964225769, + "step": 1960, + "valid_targets_mean": 2912.9, + "valid_targets_min": 1474 + }, + { + "epoch": 3.666044776119403, + "grad_norm": 0.6456842273879796, + "learning_rate": 2.1858448158369027e-05, + "loss": 0.4599, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19918210804462433, + "step": 1965, + "valid_targets_mean": 3611.8, + "valid_targets_min": 1223 + }, + { + "epoch": 3.675373134328358, + "grad_norm": 0.6586529347503346, + "learning_rate": 2.1765774344931877e-05, + "loss": 0.4606, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22834858298301697, + "step": 1970, + "valid_targets_mean": 3655.9, + "valid_targets_min": 722 + }, + { + "epoch": 3.6847014925373136, + "grad_norm": 0.658986496823165, + "learning_rate": 2.167306230457564e-05, + "loss": 0.4401, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25323599576950073, + "step": 1975, + "valid_targets_mean": 3876.9, + "valid_targets_min": 1987 + }, + { + "epoch": 3.6940298507462686, + "grad_norm": 0.6243409536030085, + "learning_rate": 2.1580314044405966e-05, + "loss": 0.4749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18794472515583038, + "step": 1980, + "valid_targets_mean": 2868.2, + "valid_targets_min": 1608 + }, + { + "epoch": 3.703358208955224, + "grad_norm": 0.6580923056546419, + "learning_rate": 2.1487531572312626e-05, + "loss": 0.4819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2263641357421875, + "step": 1985, + "valid_targets_mean": 4125.1, + "valid_targets_min": 1095 + }, + { + "epoch": 3.7126865671641793, + "grad_norm": 0.5765995738159063, + "learning_rate": 2.139471689692602e-05, + "loss": 0.4421, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23955413699150085, + "step": 1990, + "valid_targets_mean": 5533.4, + "valid_targets_min": 1430 + }, + { + "epoch": 3.7220149253731343, + "grad_norm": 0.6937898089609471, + "learning_rate": 2.130187202757375e-05, + "loss": 0.4929, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2732415199279785, + "step": 1995, + "valid_targets_mean": 3785.4, + "valid_targets_min": 1120 + }, + { + "epoch": 3.7313432835820897, + "grad_norm": 0.6172380820785784, + "learning_rate": 2.1208998974237036e-05, + "loss": 0.4878, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2721906900405884, + "step": 2000, + "valid_targets_mean": 3193.5, + "valid_targets_min": 954 + }, + { + "epoch": 3.7406716417910446, + "grad_norm": 0.7234731326467457, + "learning_rate": 2.1116099747507274e-05, + "loss": 0.4467, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2915199100971222, + "step": 2005, + "valid_targets_mean": 4389.2, + "valid_targets_min": 892 + }, + { + "epoch": 3.75, + "grad_norm": 0.7128127669753955, + "learning_rate": 2.1023176358542484e-05, + "loss": 0.4796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2387179136276245, + "step": 2010, + "valid_targets_mean": 3294.0, + "valid_targets_min": 1109 + }, + { + "epoch": 3.7593283582089554, + "grad_norm": 0.5601550837155556, + "learning_rate": 2.093023081902375e-05, + "loss": 0.4162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21390098333358765, + "step": 2015, + "valid_targets_mean": 4924.5, + "valid_targets_min": 950 + }, + { + "epoch": 3.7686567164179103, + "grad_norm": 0.6964830013803045, + "learning_rate": 2.083726514111171e-05, + "loss": 0.4347, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29899802803993225, + "step": 2020, + "valid_targets_mean": 3776.1, + "valid_targets_min": 1405 + }, + { + "epoch": 3.7779850746268657, + "grad_norm": 0.7366888340064104, + "learning_rate": 2.0744281337402966e-05, + "loss": 0.4171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18955278396606445, + "step": 2025, + "valid_targets_mean": 3049.2, + "valid_targets_min": 1828 + }, + { + "epoch": 3.7873134328358207, + "grad_norm": 0.6936177979786382, + "learning_rate": 2.0651281420886522e-05, + "loss": 0.4405, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2837923467159271, + "step": 2030, + "valid_targets_mean": 3867.2, + "valid_targets_min": 1752 + }, + { + "epoch": 3.796641791044776, + "grad_norm": 0.5965755695200002, + "learning_rate": 2.0558267404900206e-05, + "loss": 0.4841, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.260015070438385, + "step": 2035, + "valid_targets_mean": 5319.5, + "valid_targets_min": 1796 + }, + { + "epoch": 3.8059701492537314, + "grad_norm": 0.6739962889003045, + "learning_rate": 2.0465241303087078e-05, + "loss": 0.4707, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3292674720287323, + "step": 2040, + "valid_targets_mean": 4520.4, + "valid_targets_min": 1887 + }, + { + "epoch": 3.8152985074626864, + "grad_norm": 0.6442611612015025, + "learning_rate": 2.0372205129351856e-05, + "loss": 0.4334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2398635596036911, + "step": 2045, + "valid_targets_mean": 3843.2, + "valid_targets_min": 975 + }, + { + "epoch": 3.824626865671642, + "grad_norm": 0.703147087497016, + "learning_rate": 2.0279160897817283e-05, + "loss": 0.4553, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24372057616710663, + "step": 2050, + "valid_targets_mean": 2944.0, + "valid_targets_min": 1273 + }, + { + "epoch": 3.833955223880597, + "grad_norm": 0.5823014707490349, + "learning_rate": 2.0186110622780558e-05, + "loss": 0.4481, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2016904652118683, + "step": 2055, + "valid_targets_mean": 4031.4, + "valid_targets_min": 1661 + }, + { + "epoch": 3.843283582089552, + "grad_norm": 0.8482690011627374, + "learning_rate": 2.0093056318669718e-05, + "loss": 0.4411, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1738094687461853, + "step": 2060, + "valid_targets_mean": 2652.8, + "valid_targets_min": 1245 + }, + { + "epoch": 3.8526119402985075, + "grad_norm": 0.7505408174005662, + "learning_rate": 2e-05, + "loss": 0.488, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20642517507076263, + "step": 2065, + "valid_targets_mean": 2361.2, + "valid_targets_min": 679 + }, + { + "epoch": 3.861940298507463, + "grad_norm": 0.7218451651014679, + "learning_rate": 1.990694368133029e-05, + "loss": 0.4782, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26526278257369995, + "step": 2070, + "valid_targets_mean": 3974.1, + "valid_targets_min": 1214 + }, + { + "epoch": 3.871268656716418, + "grad_norm": 0.6521019408783081, + "learning_rate": 1.981388937721945e-05, + "loss": 0.4561, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19692154228687286, + "step": 2075, + "valid_targets_mean": 3396.6, + "valid_targets_min": 1637 + }, + { + "epoch": 3.8805970149253732, + "grad_norm": 0.6400940729932223, + "learning_rate": 1.9720839102182717e-05, + "loss": 0.4576, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2274247705936432, + "step": 2080, + "valid_targets_mean": 3804.2, + "valid_targets_min": 1439 + }, + { + "epoch": 3.8899253731343286, + "grad_norm": 0.5872565798013654, + "learning_rate": 1.962779487064815e-05, + "loss": 0.4903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2871890068054199, + "step": 2085, + "valid_targets_mean": 5241.0, + "valid_targets_min": 2325 + }, + { + "epoch": 3.8992537313432836, + "grad_norm": 0.979618761172191, + "learning_rate": 1.9534758696912922e-05, + "loss": 0.4699, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24316516518592834, + "step": 2090, + "valid_targets_mean": 3387.8, + "valid_targets_min": 1025 + }, + { + "epoch": 3.908582089552239, + "grad_norm": 0.773525975406754, + "learning_rate": 1.9441732595099804e-05, + "loss": 0.4733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27585703134536743, + "step": 2095, + "valid_targets_mean": 2789.4, + "valid_targets_min": 584 + }, + { + "epoch": 3.917910447761194, + "grad_norm": 0.7637276423844814, + "learning_rate": 1.9348718579113484e-05, + "loss": 0.4907, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22049155831336975, + "step": 2100, + "valid_targets_mean": 1644.9, + "valid_targets_min": 273 + }, + { + "epoch": 3.9272388059701493, + "grad_norm": 0.6176106209901676, + "learning_rate": 1.9255718662597044e-05, + "loss": 0.426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19905486702919006, + "step": 2105, + "valid_targets_mean": 3707.4, + "valid_targets_min": 1406 + }, + { + "epoch": 3.9365671641791042, + "grad_norm": 0.5810387597470694, + "learning_rate": 1.9162734858888293e-05, + "loss": 0.4605, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2654506266117096, + "step": 2110, + "valid_targets_mean": 5372.4, + "valid_targets_min": 2038 + }, + { + "epoch": 3.9458955223880596, + "grad_norm": 0.6028047075488562, + "learning_rate": 1.9069769180976263e-05, + "loss": 0.4309, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17171043157577515, + "step": 2115, + "valid_targets_mean": 3269.9, + "valid_targets_min": 1257 + }, + { + "epoch": 3.955223880597015, + "grad_norm": 0.6450131990155004, + "learning_rate": 1.8976823641457523e-05, + "loss": 0.4691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2730379104614258, + "step": 2120, + "valid_targets_mean": 3788.6, + "valid_targets_min": 1302 + }, + { + "epoch": 3.96455223880597, + "grad_norm": 0.6555567328927122, + "learning_rate": 1.888390025249273e-05, + "loss": 0.4808, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20766517519950867, + "step": 2125, + "valid_targets_mean": 2685.5, + "valid_targets_min": 1289 + }, + { + "epoch": 3.9738805970149254, + "grad_norm": 0.6624567840059948, + "learning_rate": 1.879100102576297e-05, + "loss": 0.4525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23509281873703003, + "step": 2130, + "valid_targets_mean": 3803.6, + "valid_targets_min": 1889 + }, + { + "epoch": 3.9832089552238807, + "grad_norm": 2.4947822110040794, + "learning_rate": 1.8698127972426255e-05, + "loss": 0.4697, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2938823997974396, + "step": 2135, + "valid_targets_mean": 5112.8, + "valid_targets_min": 1091 + }, + { + "epoch": 3.9925373134328357, + "grad_norm": 0.5780907568759792, + "learning_rate": 1.8605283103073982e-05, + "loss": 0.4348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27421608567237854, + "step": 2140, + "valid_targets_mean": 5806.4, + "valid_targets_min": 2032 + }, + { + "epoch": 4.001865671641791, + "grad_norm": 0.6060632020047783, + "learning_rate": 1.851246842768738e-05, + "loss": 0.4435, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24582168459892273, + "step": 2145, + "valid_targets_mean": 3831.4, + "valid_targets_min": 1441 + }, + { + "epoch": 4.0111940298507465, + "grad_norm": 0.7009901036431029, + "learning_rate": 1.841968595559404e-05, + "loss": 0.4323, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18534894287586212, + "step": 2150, + "valid_targets_mean": 3017.4, + "valid_targets_min": 369 + }, + { + "epoch": 4.020522388059701, + "grad_norm": 0.6431801674238758, + "learning_rate": 1.8326937695424367e-05, + "loss": 0.4378, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26719096302986145, + "step": 2155, + "valid_targets_mean": 4756.9, + "valid_targets_min": 700 + }, + { + "epoch": 4.029850746268656, + "grad_norm": 0.6611686880788306, + "learning_rate": 1.8234225655068133e-05, + "loss": 0.4391, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1821727752685547, + "step": 2160, + "valid_targets_mean": 2917.9, + "valid_targets_min": 2172 + }, + { + "epoch": 4.039179104477612, + "grad_norm": 0.6835057889892572, + "learning_rate": 1.8141551841630976e-05, + "loss": 0.4292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18096674978733063, + "step": 2165, + "valid_targets_mean": 2570.5, + "valid_targets_min": 711 + }, + { + "epoch": 4.048507462686567, + "grad_norm": 0.702963636690837, + "learning_rate": 1.804891826139099e-05, + "loss": 0.4263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21073882281780243, + "step": 2170, + "valid_targets_mean": 3497.5, + "valid_targets_min": 1436 + }, + { + "epoch": 4.057835820895522, + "grad_norm": 0.5936413868801393, + "learning_rate": 1.795632691975526e-05, + "loss": 0.442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10570041835308075, + "step": 2175, + "valid_targets_mean": 2645.6, + "valid_targets_min": 1344 + }, + { + "epoch": 4.067164179104478, + "grad_norm": 0.5897651964604801, + "learning_rate": 1.7863779821216443e-05, + "loss": 0.4644, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19892853498458862, + "step": 2180, + "valid_targets_mean": 4334.6, + "valid_targets_min": 786 + }, + { + "epoch": 4.076492537313433, + "grad_norm": 0.630359088758679, + "learning_rate": 1.77712789693094e-05, + "loss": 0.4248, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24680113792419434, + "step": 2185, + "valid_targets_mean": 5018.2, + "valid_targets_min": 839 + }, + { + "epoch": 4.085820895522388, + "grad_norm": 0.6475126024977133, + "learning_rate": 1.76788263665678e-05, + "loss": 0.4503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2277972400188446, + "step": 2190, + "valid_targets_mean": 3480.6, + "valid_targets_min": 1260 + }, + { + "epoch": 4.095149253731344, + "grad_norm": 0.6561200489136899, + "learning_rate": 1.758642401448078e-05, + "loss": 0.4511, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13427501916885376, + "step": 2195, + "valid_targets_mean": 2769.2, + "valid_targets_min": 1133 + }, + { + "epoch": 4.104477611940299, + "grad_norm": 0.6544497448636003, + "learning_rate": 1.74940739134496e-05, + "loss": 0.4513, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2530699372291565, + "step": 2200, + "valid_targets_mean": 5376.2, + "valid_targets_min": 2386 + }, + { + "epoch": 4.1138059701492535, + "grad_norm": 0.6256798516671497, + "learning_rate": 1.7401778062744365e-05, + "loss": 0.4478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22730037569999695, + "step": 2205, + "valid_targets_mean": 4033.2, + "valid_targets_min": 2698 + }, + { + "epoch": 4.123134328358209, + "grad_norm": 0.6079814510809718, + "learning_rate": 1.7309538460460698e-05, + "loss": 0.4275, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31900912523269653, + "step": 2210, + "valid_targets_mean": 5437.2, + "valid_targets_min": 1134 + }, + { + "epoch": 4.132462686567164, + "grad_norm": 0.647355724782543, + "learning_rate": 1.721735710347654e-05, + "loss": 0.437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3508025109767914, + "step": 2215, + "valid_targets_mean": 5634.6, + "valid_targets_min": 2196 + }, + { + "epoch": 4.141791044776119, + "grad_norm": 0.6152148655121742, + "learning_rate": 1.7125235987408867e-05, + "loss": 0.4579, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2936285138130188, + "step": 2220, + "valid_targets_mean": 5473.9, + "valid_targets_min": 1422 + }, + { + "epoch": 4.151119402985074, + "grad_norm": 0.6958876258775811, + "learning_rate": 1.7033177106570516e-05, + "loss": 0.4611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25055307149887085, + "step": 2225, + "valid_targets_mean": 3369.1, + "valid_targets_min": 872 + }, + { + "epoch": 4.16044776119403, + "grad_norm": 0.6469152760776189, + "learning_rate": 1.694118245392703e-05, + "loss": 0.437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18549364805221558, + "step": 2230, + "valid_targets_mean": 3206.5, + "valid_targets_min": 1250 + }, + { + "epoch": 4.169776119402985, + "grad_norm": 0.5615080382367836, + "learning_rate": 1.684925402105344e-05, + "loss": 0.4305, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1907576620578766, + "step": 2235, + "valid_targets_mean": 3946.8, + "valid_targets_min": 1445 + }, + { + "epoch": 4.17910447761194, + "grad_norm": 0.5136642593551568, + "learning_rate": 1.6757393798091245e-05, + "loss": 0.4531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24885302782058716, + "step": 2240, + "valid_targets_mean": 6832.8, + "valid_targets_min": 1913 + }, + { + "epoch": 4.188432835820896, + "grad_norm": 0.5648835716252492, + "learning_rate": 1.6665603773705247e-05, + "loss": 0.4587, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24105992913246155, + "step": 2245, + "valid_targets_mean": 4576.6, + "valid_targets_min": 1001 + }, + { + "epoch": 4.197761194029851, + "grad_norm": 0.770242929069094, + "learning_rate": 1.6573885935040554e-05, + "loss": 0.4488, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33347582817077637, + "step": 2250, + "valid_targets_mean": 3805.9, + "valid_targets_min": 1045 + }, + { + "epoch": 4.207089552238806, + "grad_norm": 0.5975461413608706, + "learning_rate": 1.648224226767952e-05, + "loss": 0.4547, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19192320108413696, + "step": 2255, + "valid_targets_mean": 3637.6, + "valid_targets_min": 1308 + }, + { + "epoch": 4.2164179104477615, + "grad_norm": 0.8518760022154768, + "learning_rate": 1.6390674755598798e-05, + "loss": 0.469, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2788088321685791, + "step": 2260, + "valid_targets_mean": 2425.1, + "valid_targets_min": 881 + }, + { + "epoch": 4.225746268656716, + "grad_norm": 0.6174032472222464, + "learning_rate": 1.6299185381126344e-05, + "loss": 0.4293, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17284691333770752, + "step": 2265, + "valid_targets_mean": 3921.8, + "valid_targets_min": 1312 + }, + { + "epoch": 4.235074626865671, + "grad_norm": 0.6661487358611796, + "learning_rate": 1.6207776124898557e-05, + "loss": 0.4397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27458280324935913, + "step": 2270, + "valid_targets_mean": 4506.2, + "valid_targets_min": 1758 + }, + { + "epoch": 4.244402985074627, + "grad_norm": 0.7443198937477352, + "learning_rate": 1.611644896581736e-05, + "loss": 0.4497, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.348906010389328, + "step": 2275, + "valid_targets_mean": 3673.1, + "valid_targets_min": 1133 + }, + { + "epoch": 4.253731343283582, + "grad_norm": 0.7514602820618991, + "learning_rate": 1.602520588100735e-05, + "loss": 0.4176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16066180169582367, + "step": 2280, + "valid_targets_mean": 2503.1, + "valid_targets_min": 1278 + }, + { + "epoch": 4.263059701492537, + "grad_norm": 0.6605663564712001, + "learning_rate": 1.5934048845773055e-05, + "loss": 0.4293, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1894732415676117, + "step": 2285, + "valid_targets_mean": 2932.2, + "valid_targets_min": 855 + }, + { + "epoch": 4.272388059701493, + "grad_norm": 0.6478329485707311, + "learning_rate": 1.5842979833556084e-05, + "loss": 0.4314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15124742686748505, + "step": 2290, + "valid_targets_mean": 2438.5, + "valid_targets_min": 346 + }, + { + "epoch": 4.281716417910448, + "grad_norm": 0.5804929237474775, + "learning_rate": 1.575200081589251e-05, + "loss": 0.3959, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2257540225982666, + "step": 2295, + "valid_targets_mean": 5449.8, + "valid_targets_min": 943 + }, + { + "epoch": 4.291044776119403, + "grad_norm": 0.7037622320049316, + "learning_rate": 1.5661113762370063e-05, + "loss": 0.4322, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15288349986076355, + "step": 2300, + "valid_targets_mean": 2822.6, + "valid_targets_min": 1210 + }, + { + "epoch": 4.300373134328359, + "grad_norm": 0.6963070697520833, + "learning_rate": 1.5570320640585624e-05, + "loss": 0.4234, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22153517603874207, + "step": 2305, + "valid_targets_mean": 3570.4, + "valid_targets_min": 1459 + }, + { + "epoch": 4.309701492537314, + "grad_norm": 0.7329729659703779, + "learning_rate": 1.5479623416102512e-05, + "loss": 0.4109, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2383253276348114, + "step": 2310, + "valid_targets_mean": 3944.4, + "valid_targets_min": 1938 + }, + { + "epoch": 4.3190298507462686, + "grad_norm": 0.7985277868290318, + "learning_rate": 1.5389024052408026e-05, + "loss": 0.4344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19057363271713257, + "step": 2315, + "valid_targets_mean": 2628.9, + "valid_targets_min": 791 + }, + { + "epoch": 4.3283582089552235, + "grad_norm": 0.5978848093323659, + "learning_rate": 1.529852451087085e-05, + "loss": 0.444, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2259930968284607, + "step": 2320, + "valid_targets_mean": 4019.9, + "valid_targets_min": 1779 + }, + { + "epoch": 4.337686567164179, + "grad_norm": 0.6465879123431579, + "learning_rate": 1.5208126750698676e-05, + "loss": 0.4123, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26882416009902954, + "step": 2325, + "valid_targets_mean": 4287.9, + "valid_targets_min": 892 + }, + { + "epoch": 4.347014925373134, + "grad_norm": 0.7307291618176373, + "learning_rate": 1.5117832728895734e-05, + "loss": 0.4188, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1527242362499237, + "step": 2330, + "valid_targets_mean": 2229.6, + "valid_targets_min": 1264 + }, + { + "epoch": 4.356343283582089, + "grad_norm": 0.6392318314092182, + "learning_rate": 1.5027644400220435e-05, + "loss": 0.4563, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2487446665763855, + "step": 2335, + "valid_targets_mean": 4706.1, + "valid_targets_min": 1185 + }, + { + "epoch": 4.365671641791045, + "grad_norm": 0.6618981542626332, + "learning_rate": 1.4937563717143077e-05, + "loss": 0.4652, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28851789236068726, + "step": 2340, + "valid_targets_mean": 4294.0, + "valid_targets_min": 1556 + }, + { + "epoch": 4.375, + "grad_norm": 0.7529161579275341, + "learning_rate": 1.4847592629803538e-05, + "loss": 0.4236, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1825079619884491, + "step": 2345, + "valid_targets_mean": 2900.8, + "valid_targets_min": 1075 + }, + { + "epoch": 4.384328358208955, + "grad_norm": 0.6671351820082562, + "learning_rate": 1.4757733085969095e-05, + "loss": 0.4589, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22679242491722107, + "step": 2350, + "valid_targets_mean": 3741.5, + "valid_targets_min": 839 + }, + { + "epoch": 4.393656716417911, + "grad_norm": 0.6687639369769314, + "learning_rate": 1.4667987030992223e-05, + "loss": 0.4871, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.184140145778656, + "step": 2355, + "valid_targets_mean": 3491.0, + "valid_targets_min": 1442 + }, + { + "epoch": 4.402985074626866, + "grad_norm": 0.6199553070050913, + "learning_rate": 1.4578356407768524e-05, + "loss": 0.4555, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2220112383365631, + "step": 2360, + "valid_targets_mean": 5140.2, + "valid_targets_min": 2142 + }, + { + "epoch": 4.412313432835821, + "grad_norm": 0.6365687259670468, + "learning_rate": 1.4488843156694614e-05, + "loss": 0.4184, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17575378715991974, + "step": 2365, + "valid_targets_mean": 3332.8, + "valid_targets_min": 1449 + }, + { + "epoch": 4.4216417910447765, + "grad_norm": 0.6341053111063015, + "learning_rate": 1.4399449215626154e-05, + "loss": 0.4224, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33901840448379517, + "step": 2370, + "valid_targets_mean": 4009.6, + "valid_targets_min": 1339 + }, + { + "epoch": 4.4309701492537314, + "grad_norm": 0.6265355653881179, + "learning_rate": 1.4310176519835885e-05, + "loss": 0.4373, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1486091911792755, + "step": 2375, + "valid_targets_mean": 3080.8, + "valid_targets_min": 396 + }, + { + "epoch": 4.440298507462686, + "grad_norm": 0.6629653626197524, + "learning_rate": 1.422102700197172e-05, + "loss": 0.4314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19583386182785034, + "step": 2380, + "valid_targets_mean": 3179.4, + "valid_targets_min": 948 + }, + { + "epoch": 4.449626865671641, + "grad_norm": 0.9408179448519994, + "learning_rate": 1.4132002592014939e-05, + "loss": 0.4314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2807454466819763, + "step": 2385, + "valid_targets_mean": 3273.6, + "valid_targets_min": 860 + }, + { + "epoch": 4.458955223880597, + "grad_norm": 0.6216175434291872, + "learning_rate": 1.404310521723836e-05, + "loss": 0.4533, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23090285062789917, + "step": 2390, + "valid_targets_mean": 4614.0, + "valid_targets_min": 679 + }, + { + "epoch": 4.468283582089552, + "grad_norm": 0.6226975452937478, + "learning_rate": 1.3954336802164658e-05, + "loss": 0.4635, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1825587898492813, + "step": 2395, + "valid_targets_mean": 3619.8, + "valid_targets_min": 721 + }, + { + "epoch": 4.477611940298507, + "grad_norm": 0.7638470348141475, + "learning_rate": 1.3865699268524668e-05, + "loss": 0.459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2274436205625534, + "step": 2400, + "valid_targets_mean": 3343.1, + "valid_targets_min": 1134 + }, + { + "epoch": 4.486940298507463, + "grad_norm": 0.7122301912413254, + "learning_rate": 1.3777194535215812e-05, + "loss": 0.4584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31319156289100647, + "step": 2405, + "valid_targets_mean": 4563.1, + "valid_targets_min": 1110 + }, + { + "epoch": 4.496268656716418, + "grad_norm": 0.7567993962695424, + "learning_rate": 1.3688824518260528e-05, + "loss": 0.4226, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13031692802906036, + "step": 2410, + "valid_targets_mean": 1921.2, + "valid_targets_min": 1227 + }, + { + "epoch": 4.505597014925373, + "grad_norm": 0.6711336883898671, + "learning_rate": 1.3600591130764825e-05, + "loss": 0.4459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18896600604057312, + "step": 2415, + "valid_targets_mean": 3321.0, + "valid_targets_min": 961 + }, + { + "epoch": 4.514925373134329, + "grad_norm": 0.6972431825434895, + "learning_rate": 1.3512496282876813e-05, + "loss": 0.4677, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3178806006908417, + "step": 2420, + "valid_targets_mean": 3565.5, + "valid_targets_min": 649 + }, + { + "epoch": 4.524253731343284, + "grad_norm": 0.6705103620682652, + "learning_rate": 1.3424541881745425e-05, + "loss": 0.4517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24133093655109406, + "step": 2425, + "valid_targets_mean": 3670.8, + "valid_targets_min": 360 + }, + { + "epoch": 4.5335820895522385, + "grad_norm": 0.6083320581699869, + "learning_rate": 1.3336729831479062e-05, + "loss": 0.4431, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23500260710716248, + "step": 2430, + "valid_targets_mean": 4631.4, + "valid_targets_min": 1012 + }, + { + "epoch": 4.542910447761194, + "grad_norm": 0.6381415679940657, + "learning_rate": 1.32490620331044e-05, + "loss": 0.4366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21626059710979462, + "step": 2435, + "valid_targets_mean": 3962.9, + "valid_targets_min": 1244 + }, + { + "epoch": 4.552238805970149, + "grad_norm": 0.6108744573121613, + "learning_rate": 1.3161540384525244e-05, + "loss": 0.4321, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16019883751869202, + "step": 2440, + "valid_targets_mean": 3793.0, + "valid_targets_min": 1181 + }, + { + "epoch": 4.561567164179104, + "grad_norm": 0.7680112408521977, + "learning_rate": 1.3074166780481412e-05, + "loss": 0.465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27843910455703735, + "step": 2445, + "valid_targets_mean": 3369.8, + "valid_targets_min": 1656 + }, + { + "epoch": 4.57089552238806, + "grad_norm": 0.7125575062540254, + "learning_rate": 1.2986943112507753e-05, + "loss": 0.436, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2446242868900299, + "step": 2450, + "valid_targets_mean": 3816.4, + "valid_targets_min": 1919 + }, + { + "epoch": 4.580223880597015, + "grad_norm": 0.6490484459473351, + "learning_rate": 1.2899871268893161e-05, + "loss": 0.445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21625559031963348, + "step": 2455, + "valid_targets_mean": 3832.1, + "valid_targets_min": 1701 + }, + { + "epoch": 4.58955223880597, + "grad_norm": 0.7288896293384832, + "learning_rate": 1.2812953134639725e-05, + "loss": 0.4428, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21461918950080872, + "step": 2460, + "valid_targets_mean": 2627.2, + "valid_targets_min": 1238 + }, + { + "epoch": 4.598880597014926, + "grad_norm": 0.6973433634351924, + "learning_rate": 1.27261905914219e-05, + "loss": 0.4047, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23721319437026978, + "step": 2465, + "valid_targets_mean": 4540.2, + "valid_targets_min": 311 + }, + { + "epoch": 4.608208955223881, + "grad_norm": 0.6063446914924041, + "learning_rate": 1.2639585517545783e-05, + "loss": 0.4443, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32490983605384827, + "step": 2470, + "valid_targets_mean": 6601.0, + "valid_targets_min": 1059 + }, + { + "epoch": 4.617537313432836, + "grad_norm": 0.629631304202207, + "learning_rate": 1.2553139787908457e-05, + "loss": 0.4293, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19959989190101624, + "step": 2475, + "valid_targets_mean": 4318.9, + "valid_targets_min": 770 + }, + { + "epoch": 4.6268656716417915, + "grad_norm": 0.7008850064972045, + "learning_rate": 1.2466855273957372e-05, + "loss": 0.4495, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24872657656669617, + "step": 2480, + "valid_targets_mean": 3817.8, + "valid_targets_min": 1786 + }, + { + "epoch": 4.6361940298507465, + "grad_norm": 0.7094606616520585, + "learning_rate": 1.2380733843649866e-05, + "loss": 0.4221, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17468532919883728, + "step": 2485, + "valid_targets_mean": 2549.5, + "valid_targets_min": 1984 + }, + { + "epoch": 4.645522388059701, + "grad_norm": 0.642921832086949, + "learning_rate": 1.2294777361412696e-05, + "loss": 0.4318, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22473353147506714, + "step": 2490, + "valid_targets_mean": 4510.4, + "valid_targets_min": 1683 + }, + { + "epoch": 4.654850746268656, + "grad_norm": 0.5861440734323993, + "learning_rate": 1.2208987688101701e-05, + "loss": 0.4012, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14747220277786255, + "step": 2495, + "valid_targets_mean": 3801.1, + "valid_targets_min": 1819 + }, + { + "epoch": 4.664179104477612, + "grad_norm": 0.6963996311121602, + "learning_rate": 1.2123366680961493e-05, + "loss": 0.4257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1983906626701355, + "step": 2500, + "valid_targets_mean": 3466.6, + "valid_targets_min": 827 + }, + { + "epoch": 4.673507462686567, + "grad_norm": 0.7725469692070901, + "learning_rate": 1.2037916193585285e-05, + "loss": 0.4366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20175820589065552, + "step": 2505, + "valid_targets_mean": 2666.6, + "valid_targets_min": 981 + }, + { + "epoch": 4.682835820895522, + "grad_norm": 0.7198799605934142, + "learning_rate": 1.1952638075874696e-05, + "loss": 0.4135, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13449133932590485, + "step": 2510, + "valid_targets_mean": 1679.4, + "valid_targets_min": 1285 + }, + { + "epoch": 4.692164179104478, + "grad_norm": 0.663444652586058, + "learning_rate": 1.18675341739998e-05, + "loss": 0.4314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23156410455703735, + "step": 2515, + "valid_targets_mean": 4048.6, + "valid_targets_min": 1849 + }, + { + "epoch": 4.701492537313433, + "grad_norm": 0.6538972343971937, + "learning_rate": 1.178260633035906e-05, + "loss": 0.4141, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1897672414779663, + "step": 2520, + "valid_targets_mean": 2916.8, + "valid_targets_min": 975 + }, + { + "epoch": 4.710820895522388, + "grad_norm": 0.5917015740244828, + "learning_rate": 1.1697856383539507e-05, + "loss": 0.4347, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1782170534133911, + "step": 2525, + "valid_targets_mean": 2594.5, + "valid_targets_min": 988 + }, + { + "epoch": 4.720149253731344, + "grad_norm": 0.6821840942453192, + "learning_rate": 1.1613286168276933e-05, + "loss": 0.4327, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.162733256816864, + "step": 2530, + "valid_targets_mean": 2912.6, + "valid_targets_min": 1497 + }, + { + "epoch": 4.729477611940299, + "grad_norm": 0.6424322417514464, + "learning_rate": 1.1528897515416114e-05, + "loss": 0.4533, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22855347394943237, + "step": 2535, + "valid_targets_mean": 4387.5, + "valid_targets_min": 1914 + }, + { + "epoch": 4.7388059701492535, + "grad_norm": 0.7776055816892727, + "learning_rate": 1.1444692251871253e-05, + "loss": 0.4368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1997443288564682, + "step": 2540, + "valid_targets_mean": 3196.5, + "valid_targets_min": 699 + }, + { + "epoch": 4.7481343283582085, + "grad_norm": 0.6715818075836522, + "learning_rate": 1.1360672200586365e-05, + "loss": 0.4402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26281169056892395, + "step": 2545, + "valid_targets_mean": 4101.2, + "valid_targets_min": 2035 + }, + { + "epoch": 4.757462686567164, + "grad_norm": 0.7928535784200762, + "learning_rate": 1.1276839180495877e-05, + "loss": 0.4319, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3162769377231598, + "step": 2550, + "valid_targets_mean": 3788.5, + "valid_targets_min": 1009 + }, + { + "epoch": 4.766791044776119, + "grad_norm": 0.7013727566266669, + "learning_rate": 1.1193195006485153e-05, + "loss": 0.4603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2292349934577942, + "step": 2555, + "valid_targets_mean": 3295.5, + "valid_targets_min": 955 + }, + { + "epoch": 4.776119402985074, + "grad_norm": 0.5817888769030757, + "learning_rate": 1.1109741489351325e-05, + "loss": 0.4372, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21150292456150055, + "step": 2560, + "valid_targets_mean": 4384.8, + "valid_targets_min": 1099 + }, + { + "epoch": 4.78544776119403, + "grad_norm": 0.7884954979754613, + "learning_rate": 1.1026480435763984e-05, + "loss": 0.4426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24815651774406433, + "step": 2565, + "valid_targets_mean": 2872.6, + "valid_targets_min": 745 + }, + { + "epoch": 4.794776119402985, + "grad_norm": 0.6940086966769629, + "learning_rate": 1.0943413648226142e-05, + "loss": 0.4399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19736522436141968, + "step": 2570, + "valid_targets_mean": 3237.1, + "valid_targets_min": 1805 + }, + { + "epoch": 4.80410447761194, + "grad_norm": 0.6527719590595376, + "learning_rate": 1.0860542925035187e-05, + "loss": 0.4358, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1981087327003479, + "step": 2575, + "valid_targets_mean": 3470.6, + "valid_targets_min": 1516 + }, + { + "epoch": 4.813432835820896, + "grad_norm": 0.7212961265922682, + "learning_rate": 1.0777870060243902e-05, + "loss": 0.4645, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14289066195487976, + "step": 2580, + "valid_targets_mean": 2402.8, + "valid_targets_min": 808 + }, + { + "epoch": 4.822761194029851, + "grad_norm": 0.6771507105035453, + "learning_rate": 1.0695396843621722e-05, + "loss": 0.4278, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2371062934398651, + "step": 2585, + "valid_targets_mean": 4137.4, + "valid_targets_min": 1394 + }, + { + "epoch": 4.832089552238806, + "grad_norm": 0.7278376667899698, + "learning_rate": 1.0613125060615899e-05, + "loss": 0.4091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12925156950950623, + "step": 2590, + "valid_targets_mean": 1610.5, + "valid_targets_min": 514 + }, + { + "epoch": 4.8414179104477615, + "grad_norm": 0.6501789276367153, + "learning_rate": 1.0531056492312914e-05, + "loss": 0.4332, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14594261348247528, + "step": 2595, + "valid_targets_mean": 2760.6, + "valid_targets_min": 1324 + }, + { + "epoch": 4.850746268656716, + "grad_norm": 0.633103618407971, + "learning_rate": 1.0449192915399878e-05, + "loss": 0.4257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11315122991800308, + "step": 2600, + "valid_targets_mean": 2296.2, + "valid_targets_min": 1041 + }, + { + "epoch": 4.860074626865671, + "grad_norm": 0.6844217555712022, + "learning_rate": 1.0367536102126072e-05, + "loss": 0.4538, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15750224888324738, + "step": 2605, + "valid_targets_mean": 2542.2, + "valid_targets_min": 602 + }, + { + "epoch": 4.869402985074627, + "grad_norm": 0.7269932487235211, + "learning_rate": 1.0286087820264603e-05, + "loss": 0.3927, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18912087380886078, + "step": 2610, + "valid_targets_mean": 2959.6, + "valid_targets_min": 1321 + }, + { + "epoch": 4.878731343283582, + "grad_norm": 0.6029535795525097, + "learning_rate": 1.0204849833074122e-05, + "loss": 0.4213, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2072172909975052, + "step": 2615, + "valid_targets_mean": 4930.1, + "valid_targets_min": 1433 + }, + { + "epoch": 4.888059701492537, + "grad_norm": 0.7012326292060802, + "learning_rate": 1.0123823899260634e-05, + "loss": 0.4514, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26750418543815613, + "step": 2620, + "valid_targets_mean": 3680.5, + "valid_targets_min": 1050 + }, + { + "epoch": 4.897388059701493, + "grad_norm": 0.5743499232135608, + "learning_rate": 1.0043011772939466e-05, + "loss": 0.4231, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22266338765621185, + "step": 2625, + "valid_targets_mean": 3795.2, + "valid_targets_min": 839 + }, + { + "epoch": 4.906716417910448, + "grad_norm": 0.6400543211883544, + "learning_rate": 9.962415203597245e-06, + "loss": 0.4366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18735265731811523, + "step": 2630, + "valid_targets_mean": 3965.0, + "valid_targets_min": 1881 + }, + { + "epoch": 4.916044776119403, + "grad_norm": 0.6534979492680628, + "learning_rate": 9.882035936054042e-06, + "loss": 0.4467, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19820962846279144, + "step": 2635, + "valid_targets_mean": 3497.4, + "valid_targets_min": 924 + }, + { + "epoch": 4.925373134328359, + "grad_norm": 0.6598445501360165, + "learning_rate": 9.801875710425625e-06, + "loss": 0.4416, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24448105692863464, + "step": 2640, + "valid_targets_mean": 3618.5, + "valid_targets_min": 1971 + }, + { + "epoch": 4.934701492537314, + "grad_norm": 0.6786663732925825, + "learning_rate": 9.72193626208574e-06, + "loss": 0.4422, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28025513887405396, + "step": 2645, + "valid_targets_mean": 4469.2, + "valid_targets_min": 2162 + }, + { + "epoch": 4.9440298507462686, + "grad_norm": 0.6719035268513569, + "learning_rate": 9.642219321628592e-06, + "loss": 0.4369, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23438382148742676, + "step": 2650, + "valid_targets_mean": 4588.0, + "valid_targets_min": 1468 + }, + { + "epoch": 4.9533582089552235, + "grad_norm": 0.6850595903904508, + "learning_rate": 9.56272661483133e-06, + "loss": 0.4494, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2651742696762085, + "step": 2655, + "valid_targets_mean": 4798.2, + "valid_targets_min": 1018 + }, + { + "epoch": 4.962686567164179, + "grad_norm": 0.6506478900163718, + "learning_rate": 9.483459862616723e-06, + "loss": 0.4403, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3017311096191406, + "step": 2660, + "valid_targets_mean": 5628.1, + "valid_targets_min": 1479 + }, + { + "epoch": 4.972014925373134, + "grad_norm": 0.6580283309162392, + "learning_rate": 9.404420781015881e-06, + "loss": 0.4161, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15099717676639557, + "step": 2665, + "valid_targets_mean": 2355.9, + "valid_targets_min": 1215 + }, + { + "epoch": 4.981343283582089, + "grad_norm": 0.673732605625557, + "learning_rate": 9.325611081131129e-06, + "loss": 0.4598, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2350398451089859, + "step": 2670, + "valid_targets_mean": 3777.8, + "valid_targets_min": 2000 + }, + { + "epoch": 4.990671641791045, + "grad_norm": 0.5980676015234843, + "learning_rate": 9.247032469098953e-06, + "loss": 0.445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.244520902633667, + "step": 2675, + "valid_targets_mean": 4462.0, + "valid_targets_min": 1554 + }, + { + "epoch": 5.0, + "grad_norm": 0.7432796098004715, + "learning_rate": 9.168686646053046e-06, + "loss": 0.4368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17967258393764496, + "step": 2680, + "valid_targets_mean": 3088.5, + "valid_targets_min": 1533 + }, + { + "epoch": 5.009328358208955, + "grad_norm": 0.5650320748953177, + "learning_rate": 9.090575308087502e-06, + "loss": 0.4202, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19163274765014648, + "step": 2685, + "valid_targets_mean": 4906.4, + "valid_targets_min": 2843 + }, + { + "epoch": 5.018656716417911, + "grad_norm": 0.6338297619547821, + "learning_rate": 9.012700146220082e-06, + "loss": 0.4243, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21268635988235474, + "step": 2690, + "valid_targets_mean": 4402.0, + "valid_targets_min": 1712 + }, + { + "epoch": 5.027985074626866, + "grad_norm": 0.7896758184248664, + "learning_rate": 8.935062846355642e-06, + "loss": 0.4388, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2253761738538742, + "step": 2695, + "valid_targets_mean": 2753.4, + "valid_targets_min": 1703 + }, + { + "epoch": 5.037313432835821, + "grad_norm": 0.69177340095553, + "learning_rate": 8.857665089249572e-06, + "loss": 0.3896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17261390388011932, + "step": 2700, + "valid_targets_mean": 2918.8, + "valid_targets_min": 1735 + }, + { + "epoch": 5.0466417910447765, + "grad_norm": 0.7469894097676494, + "learning_rate": 8.780508550471483e-06, + "loss": 0.4114, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.226849764585495, + "step": 2705, + "valid_targets_mean": 3747.6, + "valid_targets_min": 868 + }, + { + "epoch": 5.0559701492537314, + "grad_norm": 0.7722742088916789, + "learning_rate": 8.703594900368866e-06, + "loss": 0.4495, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19612279534339905, + "step": 2710, + "valid_targets_mean": 2987.5, + "valid_targets_min": 1266 + }, + { + "epoch": 5.065298507462686, + "grad_norm": 0.7065452567706656, + "learning_rate": 8.62692580403097e-06, + "loss": 0.4394, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1994909644126892, + "step": 2715, + "valid_targets_mean": 2953.9, + "valid_targets_min": 942 + }, + { + "epoch": 5.074626865671641, + "grad_norm": 0.6654490490254981, + "learning_rate": 8.550502921252742e-06, + "loss": 0.3963, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11040803790092468, + "step": 2720, + "valid_targets_mean": 2099.9, + "valid_targets_min": 839 + }, + { + "epoch": 5.083955223880597, + "grad_norm": 0.6754730118320539, + "learning_rate": 8.474327906498905e-06, + "loss": 0.4255, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24235710501670837, + "step": 2725, + "valid_targets_mean": 4367.6, + "valid_targets_min": 692 + }, + { + "epoch": 5.093283582089552, + "grad_norm": 0.5957275510589598, + "learning_rate": 8.398402408868145e-06, + "loss": 0.3935, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1836654245853424, + "step": 2730, + "valid_targets_mean": 4766.2, + "valid_targets_min": 1110 + }, + { + "epoch": 5.102611940298507, + "grad_norm": 0.6670776775881953, + "learning_rate": 8.322728072057358e-06, + "loss": 0.4395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20618875324726105, + "step": 2735, + "valid_targets_mean": 3293.4, + "valid_targets_min": 1537 + }, + { + "epoch": 5.111940298507463, + "grad_norm": 0.6500446779098641, + "learning_rate": 8.247306534326147e-06, + "loss": 0.4127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22236144542694092, + "step": 2740, + "valid_targets_mean": 4163.9, + "valid_targets_min": 963 + }, + { + "epoch": 5.121268656716418, + "grad_norm": 0.6857954334714782, + "learning_rate": 8.172139428461292e-06, + "loss": 0.4143, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18901528418064117, + "step": 2745, + "valid_targets_mean": 3142.4, + "valid_targets_min": 978 + }, + { + "epoch": 5.130597014925373, + "grad_norm": 0.8680157637703501, + "learning_rate": 8.097228381741441e-06, + "loss": 0.4492, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2641044855117798, + "step": 2750, + "valid_targets_mean": 2668.2, + "valid_targets_min": 844 + }, + { + "epoch": 5.139925373134329, + "grad_norm": 0.776098126813042, + "learning_rate": 8.022575015901854e-06, + "loss": 0.4117, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3025306463241577, + "step": 2755, + "valid_targets_mean": 3590.1, + "valid_targets_min": 877 + }, + { + "epoch": 5.149253731343284, + "grad_norm": 0.8463062043276112, + "learning_rate": 7.948180947099309e-06, + "loss": 0.4148, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22754406929016113, + "step": 2760, + "valid_targets_mean": 2614.2, + "valid_targets_min": 1505 + }, + { + "epoch": 5.1585820895522385, + "grad_norm": 0.7390812546619375, + "learning_rate": 7.874047785877099e-06, + "loss": 0.4188, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15721000730991364, + "step": 2765, + "valid_targets_mean": 2251.0, + "valid_targets_min": 1047 + }, + { + "epoch": 5.167910447761194, + "grad_norm": 0.5590807227500687, + "learning_rate": 7.800177137130198e-06, + "loss": 0.4239, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1274961233139038, + "step": 2770, + "valid_targets_mean": 3844.9, + "valid_targets_min": 1295 + }, + { + "epoch": 5.177238805970149, + "grad_norm": 0.6845961833060672, + "learning_rate": 7.726570600070498e-06, + "loss": 0.4333, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22977977991104126, + "step": 2775, + "valid_targets_mean": 3048.2, + "valid_targets_min": 1609 + }, + { + "epoch": 5.186567164179104, + "grad_norm": 0.7144925585051675, + "learning_rate": 7.653229768192163e-06, + "loss": 0.4384, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22613228857517242, + "step": 2780, + "valid_targets_mean": 3329.1, + "valid_targets_min": 1846 + }, + { + "epoch": 5.19589552238806, + "grad_norm": 0.6733141968439991, + "learning_rate": 7.580156229237172e-06, + "loss": 0.4201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13405677676200867, + "step": 2785, + "valid_targets_mean": 2421.1, + "valid_targets_min": 1070 + }, + { + "epoch": 5.205223880597015, + "grad_norm": 0.6726833602463744, + "learning_rate": 7.507351565160918e-06, + "loss": 0.424, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1910320520401001, + "step": 2790, + "valid_targets_mean": 3763.1, + "valid_targets_min": 1163 + }, + { + "epoch": 5.21455223880597, + "grad_norm": 0.7016446767092265, + "learning_rate": 7.434817352097994e-06, + "loss": 0.4112, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2430885136127472, + "step": 2795, + "valid_targets_mean": 4863.0, + "valid_targets_min": 1459 + }, + { + "epoch": 5.223880597014926, + "grad_norm": 0.7888976308841222, + "learning_rate": 7.36255516032802e-06, + "loss": 0.4317, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21842098236083984, + "step": 2800, + "valid_targets_mean": 2568.4, + "valid_targets_min": 907 + }, + { + "epoch": 5.233208955223881, + "grad_norm": 0.6166016582901445, + "learning_rate": 7.290566554241705e-06, + "loss": 0.4193, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19115903973579407, + "step": 2805, + "valid_targets_mean": 4269.4, + "valid_targets_min": 2463 + }, + { + "epoch": 5.242537313432836, + "grad_norm": 0.6627778432572504, + "learning_rate": 7.218853092306937e-06, + "loss": 0.4121, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2591168284416199, + "step": 2810, + "valid_targets_mean": 4634.2, + "valid_targets_min": 1480 + }, + { + "epoch": 5.251865671641791, + "grad_norm": 0.8466140837636219, + "learning_rate": 7.147416327035057e-06, + "loss": 0.449, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1265844851732254, + "step": 2815, + "valid_targets_mean": 1549.0, + "valid_targets_min": 600 + }, + { + "epoch": 5.2611940298507465, + "grad_norm": 0.7190671171195822, + "learning_rate": 7.0762578049472574e-06, + "loss": 0.4266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2343316376209259, + "step": 2820, + "valid_targets_mean": 3625.8, + "valid_targets_min": 1392 + }, + { + "epoch": 5.270522388059701, + "grad_norm": 0.6303963053220908, + "learning_rate": 7.005379066541101e-06, + "loss": 0.4259, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18870729207992554, + "step": 2825, + "valid_targets_mean": 3790.4, + "valid_targets_min": 933 + }, + { + "epoch": 5.279850746268656, + "grad_norm": 0.6137312756719183, + "learning_rate": 6.9347816462571675e-06, + "loss": 0.4341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16674146056175232, + "step": 2830, + "valid_targets_mean": 3455.1, + "valid_targets_min": 1184 + }, + { + "epoch": 5.289179104477612, + "grad_norm": 0.6724696938870595, + "learning_rate": 6.864467072445817e-06, + "loss": 0.402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2180698812007904, + "step": 2835, + "valid_targets_mean": 5510.9, + "valid_targets_min": 2781 + }, + { + "epoch": 5.298507462686567, + "grad_norm": 0.868085025237454, + "learning_rate": 6.794436867334133e-06, + "loss": 0.4126, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23925995826721191, + "step": 2840, + "valid_targets_mean": 2959.8, + "valid_targets_min": 774 + }, + { + "epoch": 5.307835820895522, + "grad_norm": 0.7727965247678154, + "learning_rate": 6.7246925469929395e-06, + "loss": 0.4502, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13370978832244873, + "step": 2845, + "valid_targets_mean": 1874.4, + "valid_targets_min": 809 + }, + { + "epoch": 5.317164179104478, + "grad_norm": 0.8904331435367575, + "learning_rate": 6.655235621304021e-06, + "loss": 0.4033, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18272532522678375, + "step": 2850, + "valid_targets_mean": 2921.0, + "valid_targets_min": 913 + }, + { + "epoch": 5.326492537313433, + "grad_norm": 0.8017933362959264, + "learning_rate": 6.5860675939273766e-06, + "loss": 0.4271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2041250765323639, + "step": 2855, + "valid_targets_mean": 3174.2, + "valid_targets_min": 1237 + }, + { + "epoch": 5.335820895522388, + "grad_norm": 0.8233747882404984, + "learning_rate": 6.517189962268733e-06, + "loss": 0.4097, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17487163841724396, + "step": 2860, + "valid_targets_mean": 2851.5, + "valid_targets_min": 2312 + }, + { + "epoch": 5.345149253731344, + "grad_norm": 0.7443897778362639, + "learning_rate": 6.448604217447067e-06, + "loss": 0.4641, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19296571612358093, + "step": 2865, + "valid_targets_mean": 3133.0, + "valid_targets_min": 777 + }, + { + "epoch": 5.354477611940299, + "grad_norm": 0.6025606420283552, + "learning_rate": 6.3803118442623565e-06, + "loss": 0.4144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17586597800254822, + "step": 2870, + "valid_targets_mean": 4149.0, + "valid_targets_min": 1223 + }, + { + "epoch": 5.3638059701492535, + "grad_norm": 0.6659343581783279, + "learning_rate": 6.312314321163449e-06, + "loss": 0.4151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1946224421262741, + "step": 2875, + "valid_targets_mean": 3293.8, + "valid_targets_min": 1537 + }, + { + "epoch": 5.373134328358209, + "grad_norm": 0.696643437457153, + "learning_rate": 6.244613120216014e-06, + "loss": 0.416, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14468196034431458, + "step": 2880, + "valid_targets_mean": 2237.8, + "valid_targets_min": 1252 + }, + { + "epoch": 5.382462686567164, + "grad_norm": 0.7235677103989777, + "learning_rate": 6.177209707070722e-06, + "loss": 0.3955, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17538125813007355, + "step": 2885, + "valid_targets_mean": 2617.5, + "valid_targets_min": 1802 + }, + { + "epoch": 5.391791044776119, + "grad_norm": 0.7755972779952659, + "learning_rate": 6.110105540931472e-06, + "loss": 0.449, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28210440278053284, + "step": 2890, + "valid_targets_mean": 3969.2, + "valid_targets_min": 1450 + }, + { + "epoch": 5.401119402985074, + "grad_norm": 0.647274765866972, + "learning_rate": 6.043302074523827e-06, + "loss": 0.4107, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15531793236732483, + "step": 2895, + "valid_targets_mean": 2801.4, + "valid_targets_min": 1133 + }, + { + "epoch": 5.41044776119403, + "grad_norm": 0.8753875579576921, + "learning_rate": 5.976800754063552e-06, + "loss": 0.4343, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16340753436088562, + "step": 2900, + "valid_targets_mean": 1916.4, + "valid_targets_min": 526 + }, + { + "epoch": 5.419776119402985, + "grad_norm": 0.7215649428484227, + "learning_rate": 5.9106030192253275e-06, + "loss": 0.4613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2129140943288803, + "step": 2905, + "valid_targets_mean": 3166.6, + "valid_targets_min": 1873 + }, + { + "epoch": 5.42910447761194, + "grad_norm": 0.6796824485403441, + "learning_rate": 5.844710303111541e-06, + "loss": 0.437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18664315342903137, + "step": 2910, + "valid_targets_mean": 3327.1, + "valid_targets_min": 1846 + }, + { + "epoch": 5.438432835820896, + "grad_norm": 0.6692318896212334, + "learning_rate": 5.7791240322213215e-06, + "loss": 0.4181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15738973021507263, + "step": 2915, + "valid_targets_mean": 3057.9, + "valid_targets_min": 1690 + }, + { + "epoch": 5.447761194029851, + "grad_norm": 0.7955814995730689, + "learning_rate": 5.7138456264195765e-06, + "loss": 0.3974, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1776963472366333, + "step": 2920, + "valid_targets_mean": 3010.6, + "valid_targets_min": 674 + }, + { + "epoch": 5.457089552238806, + "grad_norm": 0.7001004729678523, + "learning_rate": 5.648876498906335e-06, + "loss": 0.4807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15117397904396057, + "step": 2925, + "valid_targets_mean": 2731.4, + "valid_targets_min": 602 + }, + { + "epoch": 5.4664179104477615, + "grad_norm": 0.7711781611271641, + "learning_rate": 5.584218056186113e-06, + "loss": 0.4189, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19576111435890198, + "step": 2930, + "valid_targets_mean": 3041.8, + "valid_targets_min": 817 + }, + { + "epoch": 5.475746268656716, + "grad_norm": 0.6838569161947178, + "learning_rate": 5.519871698037453e-06, + "loss": 0.4279, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1883586347103119, + "step": 2935, + "valid_targets_mean": 3094.2, + "valid_targets_min": 938 + }, + { + "epoch": 5.485074626865671, + "grad_norm": 0.7624263929002456, + "learning_rate": 5.455838817482666e-06, + "loss": 0.4577, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24204497039318085, + "step": 2940, + "valid_targets_mean": 2908.0, + "valid_targets_min": 1199 + }, + { + "epoch": 5.494402985074627, + "grad_norm": 0.6335010783857147, + "learning_rate": 5.392120800757603e-06, + "loss": 0.4297, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2556833028793335, + "step": 2945, + "valid_targets_mean": 6224.2, + "valid_targets_min": 284 + }, + { + "epoch": 5.503731343283582, + "grad_norm": 0.6336287483069856, + "learning_rate": 5.328719027281728e-06, + "loss": 0.4251, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17912890017032623, + "step": 2950, + "valid_targets_mean": 3603.6, + "valid_targets_min": 1615 + }, + { + "epoch": 5.513059701492537, + "grad_norm": 0.6862044475405126, + "learning_rate": 5.265634869628182e-06, + "loss": 0.4377, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2261010706424713, + "step": 2955, + "valid_targets_mean": 4182.2, + "valid_targets_min": 1291 + }, + { + "epoch": 5.522388059701493, + "grad_norm": 0.6674400991907332, + "learning_rate": 5.20286969349413e-06, + "loss": 0.4153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23886638879776, + "step": 2960, + "valid_targets_mean": 3886.1, + "valid_targets_min": 476 + }, + { + "epoch": 5.531716417910448, + "grad_norm": 0.6314117195966904, + "learning_rate": 5.140424857671145e-06, + "loss": 0.4108, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1526910364627838, + "step": 2965, + "valid_targets_mean": 3415.9, + "valid_targets_min": 1338 + }, + { + "epoch": 5.541044776119403, + "grad_norm": 0.6940029925514981, + "learning_rate": 5.07830171401581e-06, + "loss": 0.4204, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1888255923986435, + "step": 2970, + "valid_targets_mean": 3438.6, + "valid_targets_min": 518 + }, + { + "epoch": 5.550373134328359, + "grad_norm": 0.6111753080622749, + "learning_rate": 5.016501607420476e-06, + "loss": 0.4157, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22798597812652588, + "step": 2975, + "valid_targets_mean": 4222.9, + "valid_targets_min": 1218 + }, + { + "epoch": 5.559701492537314, + "grad_norm": 0.6681967046075253, + "learning_rate": 4.955025875784098e-06, + "loss": 0.4395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24814580380916595, + "step": 2980, + "valid_targets_mean": 4181.9, + "valid_targets_min": 1642 + }, + { + "epoch": 5.5690298507462686, + "grad_norm": 0.8139703686603935, + "learning_rate": 4.893875849983323e-06, + "loss": 0.4223, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20523005723953247, + "step": 2985, + "valid_targets_mean": 2473.4, + "valid_targets_min": 1365 + }, + { + "epoch": 5.5783582089552235, + "grad_norm": 0.763592112626437, + "learning_rate": 4.833052853843636e-06, + "loss": 0.4584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34939849376678467, + "step": 2990, + "valid_targets_mean": 3709.2, + "valid_targets_min": 825 + }, + { + "epoch": 5.587686567164179, + "grad_norm": 0.6663368162729548, + "learning_rate": 4.772558204110724e-06, + "loss": 0.4107, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26237136125564575, + "step": 2995, + "valid_targets_mean": 5118.5, + "valid_targets_min": 1976 + }, + { + "epoch": 5.597014925373134, + "grad_norm": 0.806490342081118, + "learning_rate": 4.712393210421957e-06, + "loss": 0.429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20364099740982056, + "step": 3000, + "valid_targets_mean": 2631.0, + "valid_targets_min": 1324 + }, + { + "epoch": 5.606343283582089, + "grad_norm": 0.6576439211890244, + "learning_rate": 4.652559175278062e-06, + "loss": 0.4134, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2547221779823303, + "step": 3005, + "valid_targets_mean": 4836.2, + "valid_targets_min": 692 + }, + { + "epoch": 5.615671641791045, + "grad_norm": 0.6516082661061794, + "learning_rate": 4.593057394014882e-06, + "loss": 0.3873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1882113814353943, + "step": 3010, + "valid_targets_mean": 3831.9, + "valid_targets_min": 1165 + }, + { + "epoch": 5.625, + "grad_norm": 0.5574343405497224, + "learning_rate": 4.5338891547753775e-06, + "loss": 0.403, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1985384225845337, + "step": 3015, + "valid_targets_mean": 6085.6, + "valid_targets_min": 2689 + }, + { + "epoch": 5.634328358208955, + "grad_norm": 0.6610359004153935, + "learning_rate": 4.475055738481711e-06, + "loss": 0.4058, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23995891213417053, + "step": 3020, + "valid_targets_mean": 4396.9, + "valid_targets_min": 1211 + }, + { + "epoch": 5.643656716417911, + "grad_norm": 0.6910568534423158, + "learning_rate": 4.416558418807517e-06, + "loss": 0.4497, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21296477317810059, + "step": 3025, + "valid_targets_mean": 3805.8, + "valid_targets_min": 1473 + }, + { + "epoch": 5.652985074626866, + "grad_norm": 0.8047119371084426, + "learning_rate": 4.3583984621503596e-06, + "loss": 0.4126, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23835790157318115, + "step": 3030, + "valid_targets_mean": 3202.6, + "valid_targets_min": 1211 + }, + { + "epoch": 5.662313432835821, + "grad_norm": 0.8412476624117674, + "learning_rate": 4.300577127604269e-06, + "loss": 0.4234, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22557514905929565, + "step": 3035, + "valid_targets_mean": 2449.4, + "valid_targets_min": 1819 + }, + { + "epoch": 5.6716417910447765, + "grad_norm": 0.6857523392988729, + "learning_rate": 4.243095666932531e-06, + "loss": 0.4259, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13939909636974335, + "step": 3040, + "valid_targets_mean": 2928.5, + "valid_targets_min": 865 + }, + { + "epoch": 5.6809701492537314, + "grad_norm": 0.7816465741254548, + "learning_rate": 4.18595532454055e-06, + "loss": 0.4191, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20731967687606812, + "step": 3045, + "valid_targets_mean": 2922.5, + "valid_targets_min": 1529 + }, + { + "epoch": 5.690298507462686, + "grad_norm": 0.6570088327342807, + "learning_rate": 4.129157337448926e-06, + "loss": 0.4317, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2325798124074936, + "step": 3050, + "valid_targets_mean": 4718.9, + "valid_targets_min": 1328 + }, + { + "epoch": 5.699626865671641, + "grad_norm": 0.7150884645503687, + "learning_rate": 4.072702935266677e-06, + "loss": 0.4047, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20510876178741455, + "step": 3055, + "valid_targets_mean": 3591.1, + "valid_targets_min": 1186 + }, + { + "epoch": 5.708955223880597, + "grad_norm": 0.6690319491136354, + "learning_rate": 4.016593340164618e-06, + "loss": 0.4091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14075709879398346, + "step": 3060, + "valid_targets_mean": 3247.1, + "valid_targets_min": 1336 + }, + { + "epoch": 5.718283582089552, + "grad_norm": 0.8715275867533018, + "learning_rate": 3.960829766848893e-06, + "loss": 0.4141, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12968966364860535, + "step": 3065, + "valid_targets_mean": 2196.2, + "valid_targets_min": 1005 + }, + { + "epoch": 5.727611940298507, + "grad_norm": 0.7556925856308164, + "learning_rate": 3.905413422534696e-06, + "loss": 0.4296, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23165617883205414, + "step": 3070, + "valid_targets_mean": 3119.0, + "valid_targets_min": 1172 + }, + { + "epoch": 5.736940298507463, + "grad_norm": 0.6971225006979036, + "learning_rate": 3.85034550692011e-06, + "loss": 0.3867, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2653481364250183, + "step": 3075, + "valid_targets_mean": 4400.4, + "valid_targets_min": 948 + }, + { + "epoch": 5.746268656716418, + "grad_norm": 0.7501807078278229, + "learning_rate": 3.7956272121601555e-06, + "loss": 0.3884, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15901848673820496, + "step": 3080, + "valid_targets_mean": 3011.4, + "valid_targets_min": 1322 + }, + { + "epoch": 5.755597014925373, + "grad_norm": 0.919191005705221, + "learning_rate": 3.7412597228409884e-06, + "loss": 0.4404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12636013329029083, + "step": 3085, + "valid_targets_mean": 1392.0, + "valid_targets_min": 650 + }, + { + "epoch": 5.764925373134329, + "grad_norm": 0.8956530903186518, + "learning_rate": 3.687244215954222e-06, + "loss": 0.4433, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24802561104297638, + "step": 3090, + "valid_targets_mean": 2106.5, + "valid_targets_min": 302 + }, + { + "epoch": 5.774253731343284, + "grad_norm": 0.6664223416320986, + "learning_rate": 3.633581860871491e-06, + "loss": 0.4044, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29356157779693604, + "step": 3095, + "valid_targets_mean": 3856.9, + "valid_targets_min": 925 + }, + { + "epoch": 5.7835820895522385, + "grad_norm": 0.6770167915526055, + "learning_rate": 3.5802738193190954e-06, + "loss": 0.4356, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22814005613327026, + "step": 3100, + "valid_targets_mean": 4039.0, + "valid_targets_min": 1197 + }, + { + "epoch": 5.792910447761194, + "grad_norm": 0.5904332253627855, + "learning_rate": 3.5273212453528705e-06, + "loss": 0.3999, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1884804666042328, + "step": 3105, + "valid_targets_mean": 4530.6, + "valid_targets_min": 1918 + }, + { + "epoch": 5.802238805970149, + "grad_norm": 0.7570344659292785, + "learning_rate": 3.4747252853332004e-06, + "loss": 0.4334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16944462060928345, + "step": 3110, + "valid_targets_mean": 2513.9, + "valid_targets_min": 477 + }, + { + "epoch": 5.811567164179104, + "grad_norm": 0.6233705522921676, + "learning_rate": 3.4224870779002117e-06, + "loss": 0.3906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17937105894088745, + "step": 3115, + "valid_targets_mean": 4073.1, + "valid_targets_min": 1360 + }, + { + "epoch": 5.82089552238806, + "grad_norm": 0.5982485870618515, + "learning_rate": 3.3706077539490933e-06, + "loss": 0.4465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20505157113075256, + "step": 3120, + "valid_targets_mean": 4467.2, + "valid_targets_min": 1150 + }, + { + "epoch": 5.830223880597015, + "grad_norm": 0.6228514633168549, + "learning_rate": 3.3190884366056532e-06, + "loss": 0.4292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23444685339927673, + "step": 3125, + "valid_targets_mean": 4762.1, + "valid_targets_min": 1303 + }, + { + "epoch": 5.83955223880597, + "grad_norm": 0.6682707619457782, + "learning_rate": 3.2679302412019665e-06, + "loss": 0.3936, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2485828697681427, + "step": 3130, + "valid_targets_mean": 4351.9, + "valid_targets_min": 1364 + }, + { + "epoch": 5.848880597014926, + "grad_norm": 0.8097780133784603, + "learning_rate": 3.2171342752522494e-06, + "loss": 0.4604, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2296757996082306, + "step": 3135, + "valid_targets_mean": 2568.6, + "valid_targets_min": 932 + }, + { + "epoch": 5.858208955223881, + "grad_norm": 0.6048001731822701, + "learning_rate": 3.166701638428895e-06, + "loss": 0.4328, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24331319332122803, + "step": 3140, + "valid_targets_mean": 4538.8, + "valid_targets_min": 1702 + }, + { + "epoch": 5.867537313432836, + "grad_norm": 0.6534078538554506, + "learning_rate": 3.1166334225386306e-06, + "loss": 0.4053, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13204309344291687, + "step": 3145, + "valid_targets_mean": 3320.6, + "valid_targets_min": 961 + }, + { + "epoch": 5.8768656716417915, + "grad_norm": 0.7237597542981903, + "learning_rate": 3.066930711498921e-06, + "loss": 0.3972, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.176609069108963, + "step": 3150, + "valid_targets_mean": 3294.2, + "valid_targets_min": 892 + }, + { + "epoch": 5.8861940298507465, + "grad_norm": 0.6457920440796694, + "learning_rate": 3.0175945813144668e-06, + "loss": 0.4064, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22954711318016052, + "step": 3155, + "valid_targets_mean": 4949.5, + "valid_targets_min": 1958 + }, + { + "epoch": 5.895522388059701, + "grad_norm": 0.6831151650417017, + "learning_rate": 2.9686261000539484e-06, + "loss": 0.4655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22442278265953064, + "step": 3160, + "valid_targets_mean": 4283.5, + "valid_targets_min": 2245 + }, + { + "epoch": 5.904850746268656, + "grad_norm": 0.7687113784331426, + "learning_rate": 2.920026327826866e-06, + "loss": 0.4079, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2781185507774353, + "step": 3165, + "valid_targets_mean": 3618.4, + "valid_targets_min": 1549 + }, + { + "epoch": 5.914179104477612, + "grad_norm": 0.9964065037789626, + "learning_rate": 2.87179631676062e-06, + "loss": 0.4253, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1957223266363144, + "step": 3170, + "valid_targets_mean": 2006.4, + "valid_targets_min": 990 + }, + { + "epoch": 5.923507462686567, + "grad_norm": 0.7063835422090396, + "learning_rate": 2.8239371109777127e-06, + "loss": 0.4172, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23344768583774567, + "step": 3175, + "valid_targets_mean": 4792.0, + "valid_targets_min": 1394 + }, + { + "epoch": 5.932835820895522, + "grad_norm": 0.7004255800009324, + "learning_rate": 2.7764497465731487e-06, + "loss": 0.4288, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.36521095037460327, + "step": 3180, + "valid_targets_mean": 5133.1, + "valid_targets_min": 1465 + }, + { + "epoch": 5.942164179104478, + "grad_norm": 0.6275852264573771, + "learning_rate": 2.72933525159202e-06, + "loss": 0.4133, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17552903294563293, + "step": 3185, + "valid_targets_mean": 4072.9, + "valid_targets_min": 415 + }, + { + "epoch": 5.951492537313433, + "grad_norm": 0.8504858813749513, + "learning_rate": 2.6825946460072237e-06, + "loss": 0.3976, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24496161937713623, + "step": 3190, + "valid_targets_mean": 2222.5, + "valid_targets_min": 856 + }, + { + "epoch": 5.960820895522388, + "grad_norm": 0.64300173903718, + "learning_rate": 2.636228941697414e-06, + "loss": 0.4335, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2607845067977905, + "step": 3195, + "valid_targets_mean": 4499.9, + "valid_targets_min": 934 + }, + { + "epoch": 5.970149253731344, + "grad_norm": 1.0273664671571736, + "learning_rate": 2.5902391424250573e-06, + "loss": 0.3967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20763085782527924, + "step": 3200, + "valid_targets_mean": 2845.5, + "valid_targets_min": 782 + }, + { + "epoch": 5.979477611940299, + "grad_norm": 0.8253886051515185, + "learning_rate": 2.544626243814732e-06, + "loss": 0.414, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21895374357700348, + "step": 3205, + "valid_targets_mean": 2869.8, + "valid_targets_min": 725 + }, + { + "epoch": 5.9888059701492535, + "grad_norm": 0.685587775292678, + "learning_rate": 2.4993912333315605e-06, + "loss": 0.4365, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2170969545841217, + "step": 3210, + "valid_targets_mean": 3810.4, + "valid_targets_min": 1267 + }, + { + "epoch": 5.9981343283582085, + "grad_norm": 0.6510372014440976, + "learning_rate": 2.4545350902598464e-06, + "loss": 0.4341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19423387944698334, + "step": 3215, + "valid_targets_mean": 4377.9, + "valid_targets_min": 1865 + }, + { + "epoch": 6.007462686567164, + "grad_norm": 0.6957886903316935, + "learning_rate": 2.4100587856818458e-06, + "loss": 0.3946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1854865849018097, + "step": 3220, + "valid_targets_mean": 4014.5, + "valid_targets_min": 1459 + }, + { + "epoch": 6.016791044776119, + "grad_norm": 0.7375540495783114, + "learning_rate": 2.3659632824567823e-06, + "loss": 0.404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1780610829591751, + "step": 3225, + "valid_targets_mean": 2992.8, + "valid_targets_min": 1786 + }, + { + "epoch": 6.026119402985074, + "grad_norm": 0.7893155845807066, + "learning_rate": 2.322249535199965e-06, + "loss": 0.4044, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1939329355955124, + "step": 3230, + "valid_targets_mean": 2229.1, + "valid_targets_min": 1045 + }, + { + "epoch": 6.03544776119403, + "grad_norm": 0.5758679692017847, + "learning_rate": 2.278918490262143e-06, + "loss": 0.4357, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23732808232307434, + "step": 3235, + "valid_targets_mean": 5999.0, + "valid_targets_min": 1738 + }, + { + "epoch": 6.044776119402985, + "grad_norm": 0.6271375760514692, + "learning_rate": 2.235971085709028e-06, + "loss": 0.4029, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22192716598510742, + "step": 3240, + "valid_targets_mean": 4646.2, + "valid_targets_min": 1454 + }, + { + "epoch": 6.05410447761194, + "grad_norm": 0.7267021981395005, + "learning_rate": 2.193408251300948e-06, + "loss": 0.4507, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3360394835472107, + "step": 3245, + "valid_targets_mean": 4387.6, + "valid_targets_min": 978 + }, + { + "epoch": 6.063432835820896, + "grad_norm": 0.7021426609277562, + "learning_rate": 2.1512309084727724e-06, + "loss": 0.4056, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20068864524364471, + "step": 3250, + "valid_targets_mean": 3784.1, + "valid_targets_min": 1260 + }, + { + "epoch": 6.072761194029851, + "grad_norm": 0.6899633945206223, + "learning_rate": 2.1094399703139113e-06, + "loss": 0.406, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20617982745170593, + "step": 3255, + "valid_targets_mean": 4027.5, + "valid_targets_min": 1486 + }, + { + "epoch": 6.082089552238806, + "grad_norm": 0.8195486220710256, + "learning_rate": 2.068036341548585e-06, + "loss": 0.4132, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23133915662765503, + "step": 3260, + "valid_targets_mean": 3196.2, + "valid_targets_min": 2035 + }, + { + "epoch": 6.0914179104477615, + "grad_norm": 0.6811915104614523, + "learning_rate": 2.027020918516216e-06, + "loss": 0.426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17442333698272705, + "step": 3265, + "valid_targets_mean": 3753.8, + "valid_targets_min": 1462 + }, + { + "epoch": 6.100746268656716, + "grad_norm": 0.7451064908678495, + "learning_rate": 1.9863945891520474e-06, + "loss": 0.4323, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2854803800582886, + "step": 3270, + "valid_targets_mean": 4286.4, + "valid_targets_min": 1858 + }, + { + "epoch": 6.110074626865671, + "grad_norm": 0.7244269207843559, + "learning_rate": 1.9461582329678895e-06, + "loss": 0.4571, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16119667887687683, + "step": 3275, + "valid_targets_mean": 2951.9, + "valid_targets_min": 336 + }, + { + "epoch": 6.119402985074627, + "grad_norm": 0.8709750659410566, + "learning_rate": 1.9063127210331145e-06, + "loss": 0.4423, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20131686329841614, + "step": 3280, + "valid_targets_mean": 3928.1, + "valid_targets_min": 1669 + }, + { + "epoch": 6.128731343283582, + "grad_norm": 0.7429602110693045, + "learning_rate": 1.8668589159557626e-06, + "loss": 0.4205, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1660965383052826, + "step": 3285, + "valid_targets_mean": 3569.0, + "valid_targets_min": 1458 + }, + { + "epoch": 6.138059701492537, + "grad_norm": 0.7131735510261875, + "learning_rate": 1.8277976718638934e-06, + "loss": 0.4174, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1863085776567459, + "step": 3290, + "valid_targets_mean": 3340.1, + "valid_targets_min": 1279 + }, + { + "epoch": 6.147388059701493, + "grad_norm": 0.7455566163381298, + "learning_rate": 1.789129834387091e-06, + "loss": 0.418, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18644563853740692, + "step": 3295, + "valid_targets_mean": 2760.5, + "valid_targets_min": 716 + }, + { + "epoch": 6.156716417910448, + "grad_norm": 0.6711883553208501, + "learning_rate": 1.7508562406381414e-06, + "loss": 0.407, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3289915919303894, + "step": 3300, + "valid_targets_mean": 5180.6, + "valid_targets_min": 1607 + }, + { + "epoch": 6.166044776119403, + "grad_norm": 0.7220319355668634, + "learning_rate": 1.7129777191949437e-06, + "loss": 0.3783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20095688104629517, + "step": 3305, + "valid_targets_mean": 3527.5, + "valid_targets_min": 1432 + }, + { + "epoch": 6.175373134328359, + "grad_norm": 0.7233061287391501, + "learning_rate": 1.6754950900825174e-06, + "loss": 0.419, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1349385380744934, + "step": 3310, + "valid_targets_mean": 2717.4, + "valid_targets_min": 518 + }, + { + "epoch": 6.184701492537314, + "grad_norm": 0.6187130925267182, + "learning_rate": 1.6384091647553125e-06, + "loss": 0.437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19661355018615723, + "step": 3315, + "valid_targets_mean": 4145.5, + "valid_targets_min": 625 + }, + { + "epoch": 6.1940298507462686, + "grad_norm": 0.7127359933706064, + "learning_rate": 1.601720746079598e-06, + "loss": 0.3872, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1949930489063263, + "step": 3320, + "valid_targets_mean": 3383.2, + "valid_targets_min": 1912 + }, + { + "epoch": 6.2033582089552235, + "grad_norm": 0.6827825276658471, + "learning_rate": 1.5654306283161002e-06, + "loss": 0.3896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16057997941970825, + "step": 3325, + "valid_targets_mean": 3141.4, + "valid_targets_min": 1613 + }, + { + "epoch": 6.212686567164179, + "grad_norm": 0.7623156686260089, + "learning_rate": 1.5295395971028094e-06, + "loss": 0.4202, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1859741508960724, + "step": 3330, + "valid_targets_mean": 2781.8, + "valid_targets_min": 639 + }, + { + "epoch": 6.222014925373134, + "grad_norm": 0.6526555317978197, + "learning_rate": 1.4940484294379442e-06, + "loss": 0.4335, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.36455029249191284, + "step": 3335, + "valid_targets_mean": 5160.5, + "valid_targets_min": 1179 + }, + { + "epoch": 6.231343283582089, + "grad_norm": 0.6151501372350625, + "learning_rate": 1.4589578936631776e-06, + "loss": 0.4131, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1599680334329605, + "step": 3340, + "valid_targets_mean": 3947.1, + "valid_targets_min": 1558 + }, + { + "epoch": 6.240671641791045, + "grad_norm": 0.6160801970086336, + "learning_rate": 1.424268749446962e-06, + "loss": 0.3663, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21979206800460815, + "step": 3345, + "valid_targets_mean": 5793.2, + "valid_targets_min": 2661 + }, + { + "epoch": 6.25, + "grad_norm": 0.7083223219195554, + "learning_rate": 1.3899817477681056e-06, + "loss": 0.4162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17735572159290314, + "step": 3350, + "valid_targets_mean": 2859.5, + "valid_targets_min": 1029 + }, + { + "epoch": 6.259328358208955, + "grad_norm": 0.8843773396491078, + "learning_rate": 1.3560976308995065e-06, + "loss": 0.3911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1908641755580902, + "step": 3355, + "valid_targets_mean": 1929.6, + "valid_targets_min": 778 + }, + { + "epoch": 6.268656716417911, + "grad_norm": 1.1978044032534327, + "learning_rate": 1.3226171323920811e-06, + "loss": 0.4385, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22838398814201355, + "step": 3360, + "valid_targets_mean": 3202.0, + "valid_targets_min": 974 + }, + { + "epoch": 6.277985074626866, + "grad_norm": 0.7494867938435484, + "learning_rate": 1.2895409770588874e-06, + "loss": 0.3937, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24941056966781616, + "step": 3365, + "valid_targets_mean": 3614.2, + "valid_targets_min": 1176 + }, + { + "epoch": 6.287313432835821, + "grad_norm": 0.7383756430752176, + "learning_rate": 1.2568698809594437e-06, + "loss": 0.4036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1690540313720703, + "step": 3370, + "valid_targets_mean": 2442.9, + "valid_targets_min": 780 + }, + { + "epoch": 6.2966417910447765, + "grad_norm": 0.6979531207282197, + "learning_rate": 1.2246045513842008e-06, + "loss": 0.3823, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17707446217536926, + "step": 3375, + "valid_targets_mean": 4107.1, + "valid_targets_min": 721 + }, + { + "epoch": 6.3059701492537314, + "grad_norm": 0.6605500097273439, + "learning_rate": 1.1927456868392605e-06, + "loss": 0.4252, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22325918078422546, + "step": 3380, + "valid_targets_mean": 4676.5, + "valid_targets_min": 2599 + }, + { + "epoch": 6.315298507462686, + "grad_norm": 0.7878548942459243, + "learning_rate": 1.1612939770312325e-06, + "loss": 0.4473, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24905769526958466, + "step": 3385, + "valid_targets_mean": 3095.1, + "valid_targets_min": 435 + }, + { + "epoch": 6.324626865671641, + "grad_norm": 1.0521696508005067, + "learning_rate": 1.1302501028523039e-06, + "loss": 0.4836, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3051932156085968, + "step": 3390, + "valid_targets_mean": 4302.6, + "valid_targets_min": 913 + }, + { + "epoch": 6.333955223880597, + "grad_norm": 0.592831512396042, + "learning_rate": 1.0996147363655175e-06, + "loss": 0.3809, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16529172658920288, + "step": 3395, + "valid_targets_mean": 4249.2, + "valid_targets_min": 1629 + }, + { + "epoch": 6.343283582089552, + "grad_norm": 0.6227501455630098, + "learning_rate": 1.069388540790195e-06, + "loss": 0.4204, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13754741847515106, + "step": 3400, + "valid_targets_mean": 4249.1, + "valid_targets_min": 1199 + }, + { + "epoch": 6.352611940298507, + "grad_norm": 0.8382135423218251, + "learning_rate": 1.039572170487606e-06, + "loss": 0.3976, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1525755226612091, + "step": 3405, + "valid_targets_mean": 2806.8, + "valid_targets_min": 792 + }, + { + "epoch": 6.361940298507463, + "grad_norm": 0.6606022400626858, + "learning_rate": 1.0101662709467841e-06, + "loss": 0.4114, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18086276948451996, + "step": 3410, + "valid_targets_mean": 3630.5, + "valid_targets_min": 1064 + }, + { + "epoch": 6.371268656716418, + "grad_norm": 0.6558442888814625, + "learning_rate": 9.811714787705528e-07, + "loss": 0.4134, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23364867269992828, + "step": 3415, + "valid_targets_mean": 3803.8, + "valid_targets_min": 515 + }, + { + "epoch": 6.380597014925373, + "grad_norm": 0.7979016301100487, + "learning_rate": 9.525884216617575e-07, + "loss": 0.4551, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3536340594291687, + "step": 3420, + "valid_targets_mean": 3563.1, + "valid_targets_min": 1117 + }, + { + "epoch": 6.389925373134329, + "grad_norm": 0.6185553560617878, + "learning_rate": 9.244177184096603e-07, + "loss": 0.4088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1816781610250473, + "step": 3425, + "valid_targets_mean": 3587.2, + "valid_targets_min": 1196 + }, + { + "epoch": 6.399253731343284, + "grad_norm": 0.6461446700758247, + "learning_rate": 8.96659978876564e-07, + "loss": 0.4445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21228957176208496, + "step": 3430, + "valid_targets_mean": 3427.9, + "valid_targets_min": 1364 + }, + { + "epoch": 6.4085820895522385, + "grad_norm": 0.6370489379600102, + "learning_rate": 8.693158039845851e-07, + "loss": 0.4023, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.237797349691391, + "step": 3435, + "valid_targets_mean": 4330.8, + "valid_targets_min": 273 + }, + { + "epoch": 6.417910447761194, + "grad_norm": 0.6582084874859667, + "learning_rate": 8.423857857026574e-07, + "loss": 0.4292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16511529684066772, + "step": 3440, + "valid_targets_mean": 4038.5, + "valid_targets_min": 1799 + }, + { + "epoch": 6.427238805970149, + "grad_norm": 0.8035397705894686, + "learning_rate": 8.158705070337181e-07, + "loss": 0.415, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1576245278120041, + "step": 3445, + "valid_targets_mean": 2312.2, + "valid_targets_min": 1100 + }, + { + "epoch": 6.436567164179104, + "grad_norm": 0.8347501785511845, + "learning_rate": 7.897705420020863e-07, + "loss": 0.417, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20214515924453735, + "step": 3450, + "valid_targets_mean": 2022.9, + "valid_targets_min": 811 + }, + { + "epoch": 6.44589552238806, + "grad_norm": 0.6870720744607335, + "learning_rate": 7.640864556410311e-07, + "loss": 0.4281, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14587557315826416, + "step": 3455, + "valid_targets_mean": 2709.8, + "valid_targets_min": 1120 + }, + { + "epoch": 6.455223880597015, + "grad_norm": 0.6842176075206781, + "learning_rate": 7.388188039805455e-07, + "loss": 0.4097, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17586246132850647, + "step": 3460, + "valid_targets_mean": 3484.8, + "valid_targets_min": 1199 + }, + { + "epoch": 6.46455223880597, + "grad_norm": 0.718267164202874, + "learning_rate": 7.139681340352966e-07, + "loss": 0.4036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14952470362186432, + "step": 3465, + "valid_targets_mean": 2353.8, + "valid_targets_min": 790 + }, + { + "epoch": 6.473880597014926, + "grad_norm": 0.629729514067453, + "learning_rate": 6.895349837928033e-07, + "loss": 0.4073, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2846236228942871, + "step": 3470, + "valid_targets_mean": 5173.2, + "valid_targets_min": 1375 + }, + { + "epoch": 6.483208955223881, + "grad_norm": 0.7361206913365843, + "learning_rate": 6.655198822017661e-07, + "loss": 0.436, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18752285838127136, + "step": 3475, + "valid_targets_mean": 3109.5, + "valid_targets_min": 1359 + }, + { + "epoch": 6.492537313432836, + "grad_norm": 0.7704518663050081, + "learning_rate": 6.419233491606403e-07, + "loss": 0.4271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23204265534877777, + "step": 3480, + "valid_targets_mean": 3539.4, + "valid_targets_min": 1296 + }, + { + "epoch": 6.5018656716417915, + "grad_norm": 0.7225818651132002, + "learning_rate": 6.187458955063652e-07, + "loss": 0.3898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20344993472099304, + "step": 3485, + "valid_targets_mean": 3525.0, + "valid_targets_min": 1672 + }, + { + "epoch": 6.5111940298507465, + "grad_norm": 0.8828281184576846, + "learning_rate": 5.959880230033043e-07, + "loss": 0.4292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16142112016677856, + "step": 3490, + "valid_targets_mean": 2504.4, + "valid_targets_min": 1382 + }, + { + "epoch": 6.520522388059701, + "grad_norm": 0.6456211710432735, + "learning_rate": 5.736502243323894e-07, + "loss": 0.4094, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20658551156520844, + "step": 3495, + "valid_targets_mean": 4238.2, + "valid_targets_min": 854 + }, + { + "epoch": 6.529850746268656, + "grad_norm": 0.6982590352107721, + "learning_rate": 5.517329830804552e-07, + "loss": 0.3866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18700110912322998, + "step": 3500, + "valid_targets_mean": 3436.9, + "valid_targets_min": 746 + }, + { + "epoch": 6.539179104477612, + "grad_norm": 1.0416057006022554, + "learning_rate": 5.302367737297642e-07, + "loss": 0.3839, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19404514133930206, + "step": 3505, + "valid_targets_mean": 3251.4, + "valid_targets_min": 1650 + }, + { + "epoch": 6.548507462686567, + "grad_norm": 0.762369543137671, + "learning_rate": 5.091620616477366e-07, + "loss": 0.4167, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22729825973510742, + "step": 3510, + "valid_targets_mean": 3530.2, + "valid_targets_min": 1117 + }, + { + "epoch": 6.557835820895522, + "grad_norm": 0.7281638887340047, + "learning_rate": 4.88509303076885e-07, + "loss": 0.3928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21897470951080322, + "step": 3515, + "valid_targets_mean": 3942.9, + "valid_targets_min": 360 + }, + { + "epoch": 6.567164179104478, + "grad_norm": 0.6979639665788098, + "learning_rate": 4.6827894512491814e-07, + "loss": 0.4061, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20810270309448242, + "step": 3520, + "valid_targets_mean": 3859.5, + "valid_targets_min": 1070 + }, + { + "epoch": 6.576492537313433, + "grad_norm": 0.7363877294361197, + "learning_rate": 4.4847142575507706e-07, + "loss": 0.395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24719929695129395, + "step": 3525, + "valid_targets_mean": 4256.1, + "valid_targets_min": 1696 + }, + { + "epoch": 6.585820895522388, + "grad_norm": 0.7823925366060069, + "learning_rate": 4.290871737766544e-07, + "loss": 0.4427, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23075155913829803, + "step": 3530, + "valid_targets_mean": 2815.9, + "valid_targets_min": 1542 + }, + { + "epoch": 6.595149253731344, + "grad_norm": 0.7701263762503229, + "learning_rate": 4.1012660883570146e-07, + "loss": 0.4219, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13105091452598572, + "step": 3535, + "valid_targets_mean": 2105.9, + "valid_targets_min": 600 + }, + { + "epoch": 6.604477611940299, + "grad_norm": 0.5841826712400768, + "learning_rate": 3.915901414059464e-07, + "loss": 0.3875, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1549183875322342, + "step": 3540, + "valid_targets_mean": 4577.4, + "valid_targets_min": 1012 + }, + { + "epoch": 6.6138059701492535, + "grad_norm": 0.6883802843116491, + "learning_rate": 3.734781727799086e-07, + "loss": 0.389, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17763827741146088, + "step": 3545, + "valid_targets_mean": 3975.0, + "valid_targets_min": 686 + }, + { + "epoch": 6.6231343283582085, + "grad_norm": 0.7558346827586868, + "learning_rate": 3.557910950602228e-07, + "loss": 0.4182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14256131649017334, + "step": 3550, + "valid_targets_mean": 1844.1, + "valid_targets_min": 781 + }, + { + "epoch": 6.632462686567164, + "grad_norm": 0.7140695573379601, + "learning_rate": 3.3852929115112177e-07, + "loss": 0.3998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1760803759098053, + "step": 3555, + "valid_targets_mean": 3544.5, + "valid_targets_min": 1418 + }, + { + "epoch": 6.641791044776119, + "grad_norm": 0.6425073528045743, + "learning_rate": 3.2169313475018506e-07, + "loss": 0.4442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18695861101150513, + "step": 3560, + "valid_targets_mean": 3875.8, + "valid_targets_min": 1434 + }, + { + "epoch": 6.651119402985074, + "grad_norm": 0.6080700632142163, + "learning_rate": 3.0528299034021437e-07, + "loss": 0.397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18689629435539246, + "step": 3565, + "valid_targets_mean": 4561.2, + "valid_targets_min": 1317 + }, + { + "epoch": 6.66044776119403, + "grad_norm": 0.7965855791582336, + "learning_rate": 2.892992131813621e-07, + "loss": 0.4202, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20818768441677094, + "step": 3570, + "valid_targets_mean": 3581.8, + "valid_targets_min": 1205 + }, + { + "epoch": 6.669776119402985, + "grad_norm": 0.6995837658044377, + "learning_rate": 2.737421493034331e-07, + "loss": 0.4157, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17812196910381317, + "step": 3575, + "valid_targets_mean": 3287.8, + "valid_targets_min": 1237 + }, + { + "epoch": 6.67910447761194, + "grad_norm": 0.6403577918014498, + "learning_rate": 2.586121354984017e-07, + "loss": 0.4163, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23359298706054688, + "step": 3580, + "valid_targets_mean": 5413.5, + "valid_targets_min": 2057 + }, + { + "epoch": 6.688432835820896, + "grad_norm": 0.57504782980932, + "learning_rate": 2.4390949931311304e-07, + "loss": 0.3958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12046363949775696, + "step": 3585, + "valid_targets_mean": 4195.2, + "valid_targets_min": 650 + }, + { + "epoch": 6.697761194029851, + "grad_norm": 0.707324481325869, + "learning_rate": 2.296345590421889e-07, + "loss": 0.4321, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22736376523971558, + "step": 3590, + "valid_targets_mean": 3745.4, + "valid_targets_min": 1434 + }, + { + "epoch": 6.707089552238806, + "grad_norm": 0.6717599047778491, + "learning_rate": 2.157876237211487e-07, + "loss": 0.3964, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2693077027797699, + "step": 3595, + "valid_targets_mean": 4930.0, + "valid_targets_min": 1969 + }, + { + "epoch": 6.7164179104477615, + "grad_norm": 0.6964655421773875, + "learning_rate": 2.0236899311970593e-07, + "loss": 0.4094, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17000341415405273, + "step": 3600, + "valid_targets_mean": 2721.8, + "valid_targets_min": 1224 + }, + { + "epoch": 6.725746268656716, + "grad_norm": 0.8319322314464308, + "learning_rate": 1.8937895773529114e-07, + "loss": 0.4351, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21979916095733643, + "step": 3605, + "valid_targets_mean": 2926.9, + "valid_targets_min": 795 + }, + { + "epoch": 6.735074626865671, + "grad_norm": 0.7774006575688079, + "learning_rate": 1.7681779878675697e-07, + "loss": 0.4176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17645777761936188, + "step": 3610, + "valid_targets_mean": 2225.2, + "valid_targets_min": 1363 + }, + { + "epoch": 6.744402985074627, + "grad_norm": 0.6469554757014483, + "learning_rate": 1.646857882082853e-07, + "loss": 0.4092, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20928488671779633, + "step": 3615, + "valid_targets_mean": 4176.9, + "valid_targets_min": 861 + }, + { + "epoch": 6.753731343283582, + "grad_norm": 0.843973781651811, + "learning_rate": 1.5298318864350738e-07, + "loss": 0.3945, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2268897294998169, + "step": 3620, + "valid_targets_mean": 2861.4, + "valid_targets_min": 808 + }, + { + "epoch": 6.763059701492537, + "grad_norm": 0.7432435120793429, + "learning_rate": 1.417102534398196e-07, + "loss": 0.4233, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09236513823270798, + "step": 3625, + "valid_targets_mean": 2268.1, + "valid_targets_min": 466 + }, + { + "epoch": 6.772388059701493, + "grad_norm": 0.6537084944755202, + "learning_rate": 1.3086722664288787e-07, + "loss": 0.4228, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24160681664943695, + "step": 3630, + "valid_targets_mean": 4067.2, + "valid_targets_min": 990 + }, + { + "epoch": 6.781716417910448, + "grad_norm": 0.6374489952440006, + "learning_rate": 1.2045434299137403e-07, + "loss": 0.405, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2295844703912735, + "step": 3635, + "valid_targets_mean": 4980.4, + "valid_targets_min": 1200 + }, + { + "epoch": 6.791044776119403, + "grad_norm": 0.8158266773809864, + "learning_rate": 1.1047182791185551e-07, + "loss": 0.4166, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20229694247245789, + "step": 3640, + "valid_targets_mean": 3240.6, + "valid_targets_min": 724 + }, + { + "epoch": 6.800373134328359, + "grad_norm": 0.6844336180185098, + "learning_rate": 1.0091989751393583e-07, + "loss": 0.4016, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19978830218315125, + "step": 3645, + "valid_targets_mean": 4074.0, + "valid_targets_min": 1011 + }, + { + "epoch": 6.809701492537314, + "grad_norm": 0.6068688014339626, + "learning_rate": 9.179875858557285e-08, + "loss": 0.4096, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19169116020202637, + "step": 3650, + "valid_targets_mean": 4783.6, + "valid_targets_min": 1819 + }, + { + "epoch": 6.8190298507462686, + "grad_norm": 0.679718685067832, + "learning_rate": 8.310860858859793e-08, + "loss": 0.3898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16579127311706543, + "step": 3655, + "valid_targets_mean": 3271.4, + "valid_targets_min": 955 + }, + { + "epoch": 6.8283582089552235, + "grad_norm": 0.6240357816971318, + "learning_rate": 7.484963565444813e-08, + "loss": 0.4027, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1807710975408554, + "step": 3660, + "valid_targets_mean": 4356.5, + "valid_targets_min": 1683 + }, + { + "epoch": 6.837686567164179, + "grad_norm": 0.8997239043284767, + "learning_rate": 6.702201858008517e-08, + "loss": 0.4045, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19364050030708313, + "step": 3665, + "valid_targets_mean": 2621.9, + "valid_targets_min": 901 + }, + { + "epoch": 6.847014925373134, + "grad_norm": 0.6551880613285674, + "learning_rate": 5.962592682412948e-08, + "loss": 0.4256, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19388341903686523, + "step": 3670, + "valid_targets_mean": 3922.9, + "valid_targets_min": 749 + }, + { + "epoch": 6.856343283582089, + "grad_norm": 0.8873475637528511, + "learning_rate": 5.2661520503192176e-08, + "loss": 0.4592, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17050158977508545, + "step": 3675, + "valid_targets_mean": 2052.1, + "valid_targets_min": 1759 + }, + { + "epoch": 6.865671641791045, + "grad_norm": 0.6620972031783834, + "learning_rate": 4.612895038840215e-08, + "loss": 0.4153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2763952612876892, + "step": 3680, + "valid_targets_mean": 5871.6, + "valid_targets_min": 823 + }, + { + "epoch": 6.875, + "grad_norm": 0.646399810747169, + "learning_rate": 4.0028357902153204e-08, + "loss": 0.4372, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23939159512519836, + "step": 3685, + "valid_targets_mean": 3650.6, + "valid_targets_min": 1637 + }, + { + "epoch": 6.884328358208955, + "grad_norm": 0.7786476743857734, + "learning_rate": 3.435987511503314e-08, + "loss": 0.4054, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2298683375120163, + "step": 3690, + "valid_targets_mean": 3284.2, + "valid_targets_min": 1414 + }, + { + "epoch": 6.893656716417911, + "grad_norm": 0.8271915376241483, + "learning_rate": 2.9123624742963818e-08, + "loss": 0.4037, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20432712137699127, + "step": 3695, + "valid_targets_mean": 3313.1, + "valid_targets_min": 1172 + }, + { + "epoch": 6.902985074626866, + "grad_norm": 0.6673844395426864, + "learning_rate": 2.4319720144554417e-08, + "loss": 0.4379, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24143289029598236, + "step": 3700, + "valid_targets_mean": 4895.2, + "valid_targets_min": 1919 + }, + { + "epoch": 6.912313432835821, + "grad_norm": 0.7429881662051876, + "learning_rate": 1.9948265318638915e-08, + "loss": 0.4209, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24234910309314728, + "step": 3705, + "valid_targets_mean": 4462.0, + "valid_targets_min": 1537 + }, + { + "epoch": 6.9216417910447765, + "grad_norm": 0.7319193423701678, + "learning_rate": 1.6009354902024598e-08, + "loss": 0.4224, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16761994361877441, + "step": 3710, + "valid_targets_mean": 3053.9, + "valid_targets_min": 884 + }, + { + "epoch": 6.9309701492537314, + "grad_norm": 0.6834045391106603, + "learning_rate": 1.2503074167451445e-08, + "loss": 0.4284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2003517895936966, + "step": 3715, + "valid_targets_mean": 4521.0, + "valid_targets_min": 1312 + }, + { + "epoch": 6.940298507462686, + "grad_norm": 0.7102406173302587, + "learning_rate": 9.429499021731403e-09, + "loss": 0.4175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22342665493488312, + "step": 3720, + "valid_targets_mean": 4343.9, + "valid_targets_min": 935 + }, + { + "epoch": 6.949626865671641, + "grad_norm": 0.6888790932368005, + "learning_rate": 6.788696004120798e-09, + "loss": 0.4521, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18838061392307281, + "step": 3725, + "valid_targets_mean": 3470.2, + "valid_targets_min": 1640 + }, + { + "epoch": 6.958955223880597, + "grad_norm": 0.7791224075152483, + "learning_rate": 4.580722284872608e-09, + "loss": 0.3958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2523302137851715, + "step": 3730, + "valid_targets_mean": 3751.1, + "valid_targets_min": 2557 + }, + { + "epoch": 6.968283582089552, + "grad_norm": 0.6981227296720393, + "learning_rate": 2.8056256639974467e-09, + "loss": 0.3764, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1694352775812149, + "step": 3735, + "valid_targets_mean": 4078.9, + "valid_targets_min": 1586 + }, + { + "epoch": 6.977611940298507, + "grad_norm": 0.6041985239305606, + "learning_rate": 1.4634445702288446e-09, + "loss": 0.4403, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17324933409690857, + "step": 3740, + "valid_targets_mean": 4397.9, + "valid_targets_min": 1706 + }, + { + "epoch": 6.986940298507463, + "grad_norm": 0.8751160509231176, + "learning_rate": 5.542080601950161e-10, + "loss": 0.3982, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1973869502544403, + "step": 3745, + "valid_targets_mean": 1969.9, + "valid_targets_min": 770 + }, + { + "epoch": 6.996268656716418, + "grad_norm": 0.5874782019411326, + "learning_rate": 7.793581778825499e-11, + "loss": 0.4035, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2725503444671631, + "step": 3750, + "valid_targets_mean": 5778.5, + "valid_targets_min": 2086 + }, + { + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.312700480222702, + "step": 3752, + "total_flos": 9.579822550267986e+17, + "train_loss": 0.48628626106136136, + "train_runtime": 32410.6521, + "train_samples_per_second": 1.851, + "train_steps_per_second": 0.116, + "valid_targets_mean": 4984.9, + "valid_targets_min": 1323 + } + ], + "logging_steps": 5, + "max_steps": 3752, + "num_input_tokens_seen": 0, + "num_train_epochs": 7, + "save_steps": 600, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 9.579822550267986e+17, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..3520b0b --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff2a5077720c82e6fdc865703a21fa05dc070084dc3cc7b4c1e5f79449bd1d13 +size 8657 diff --git a/training_loss.png b/training_loss.png new file mode 100644 index 0000000..745d7a1 Binary files /dev/null and b/training_loss.png differ diff --git a/vocab.json b/vocab.json new file mode 100644 index 0000000..6c49fc6 --- /dev/null +++ b/vocab.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910 +size 2776833